Full download text book at textbookfull.

com

Quantitative Data Analysis: A Companion

for Accounting and Information Systems
Research 1st Edition Willem Mertens
DOWLOAD HERE

https://textbookfull.com/product/quantitative-
data-analysis-a-companion-for-accounting-and-
information-systems-research-1st-edition-willem-
mertens/

DOWLOAD NOW

Download more textbook from textbookfull.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...

The Routledge Companion to Accounting Information

Systems First Edition. Edition Quinn

https://textbookfull.com/product/the-routledge-companion-to-
accounting-information-systems-first-edition-edition-quinn/

Research and Evaluation in Education and Psychology

Integrating Diversity with Quantitative Qualitative and
Mixed Methods Donna M. Mertens

https://textbookfull.com/product/research-and-evaluation-in-
education-and-psychology-integrating-diversity-with-quantitative-
qualitative-and-mixed-methods-donna-m-mertens/

Quantitative Analysis of Market Data 1st Edition Adam

Grimes

https://textbookfull.com/product/quantitative-analysis-of-market-
data-1st-edition-adam-grimes/

The Routledge Companion to Behavioral Accounting

Research 1st Edition Theresa Libby (Editor)

https://textbookfull.com/product/the-routledge-companion-to-
behavioral-accounting-research-1st-edition-theresa-libby-editor/
Quantitative Psychological Research The Complete
Student s Companion David Clark-Carter

https://textbookfull.com/product/quantitative-psychological-
research-the-complete-student-s-companion-david-clark-carter/

Quantitative Psychological Research; The Complete

Student's Companion; 5th Edition David Clark-Carter

https://textbookfull.com/product/quantitative-psychological-
research-the-complete-students-companion-5th-edition-david-clark-
carter/

Quantitative Methods of Data Analysis for the Physical

Sciences and Engineering 1st Edition Douglas G.
Martinson

https://textbookfull.com/product/quantitative-methods-of-data-
analysis-for-the-physical-sciences-and-engineering-1st-edition-
douglas-g-martinson/

Quantitative Data Analysis for Language Assessment

Volume II Advanced Methods Vahid Aryadoust

https://textbookfull.com/product/quantitative-data-analysis-for-
language-assessment-volume-ii-advanced-methods-vahid-aryadoust/

Accounting Information Systems 11th Edition Ulric J.

Gelinas

https://textbookfull.com/product/accounting-information-
systems-11th-edition-ulric-j-gelinas/
Willem Mertens · Amedeo Pugliese
Jan Recker

Quantitative
Data
Analysis
A Companion for Accounting and
Information Systems Research
Quantitative Data Analysis
Willem Mertens • Amedeo Pugliese •
Jan Recker

Quantitative Data Analysis

A Companion for Accounting and
Information Systems Research
Willem Mertens Amedeo Pugliese
QUT Business School Dept. of Economics and Management
Queensland University of Technology University of Padova
Brisbane, Queensland Padova, Italy
Australia
School of Accountancy
Queensland University of Technology
Brisbane, Queensland
Australia

Jan Recker
QUT Business School
Queensland University of Technology
Brisbane, Queensland
Australia

ISBN 978-3-319-42699-0 ISBN 978-3-319-42700-3 (eBook)

DOI 10.1007/978-3-319-42700-3

Library of Congress Control Number: 2016954196

# Springer International Publishing Switzerland 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG Switzerland
Preface

Quantitative Data Analysis for Accounting and Information Systems Research

guides postgraduate research students and early career researchers in choosing
and executing appropriate data analysis methods to answer their research questions.
It supports researchers when they are planning to collect data and when they have
data that they want to analyze. Using a variety of examples and lay language, this
book provides hands-on guidelines on (1) when to use which data analysis method,
(2) what kind of data and data structure are required for each of the methods,
(3) what each method does and how the methods should be used, and (4) how to
report results.
This book is not intended to provide an exhaustive overview of data-analysis
methods, nor does it explain the methods in depth. Instead, it guides researchers in
applying the right methods in the right way. More skilled researchers can also use
the book to refresh their knowledge or as a checklist to avoid skipping important
steps. It explains the most commonly used methods in an intuitive and hands-on
way, pointing out more advanced resources along the way. As such, it does not
aspire to compete with manuals like those of Stevens [1], Field [2], or Crawley [3].
We are not statisticians but researchers who apply statistics,1 so the book covers the
issues that commonly affect others like us, who are engaging in quantitative
empirical research.
Quantitative Data Analysis for Accounting and Information Systems Research is
the book we would have liked to have had as a support in our own research. Every
chapter provides an unintimidating starting point for building your data-analysis
skills, the information required to run the most common analyses and report them,
and pointers to more extensive resources. At the risk of saying things that may not
be entirely true in the purest statistical sense, we try to keep the language of this
book as simple as possible. As such, the book is brief and written in a language that
we hope everyone can understand—from students to researchers to people who
wish to study the organizations in which they work. Our goal is to help you conduct
academic research of high quality and do the right things right—not to make you a

1
If you are a statistician or simply more observant than we are, we invite you to tell us if you
identify an error.

v
vi Preface

statistics expert—so this book is not about statistics but about applying statistics to
the research questions that keep you awake at night. (We doubt these questions are
about collinearity, but if they are, this may not be the book you are looking for.)
In brief, this book is a software-independent starting point for answering the
question: What methods do I use to answer my research questions and how?
We hope you have fun!

Brisbane, QLD, Australia Willem Mertens

Padova, Italy Amedeo Pugliese
Jan Recker

References
1. Stevens JP (2009) Applied multivariate statistics for the social sciences. Taylor and Francis,
LLC, London
2. Field AP (2013) Discovering statistics using IBM SPSS statistics, and sex and drugs and rock
‘n’ roll, 4th edn. Sage, London
3. Crawley MJ (2013) The R book, 2nd edn. Wiley, West Sussex
Acknowledgments

Many people have contributed to the development of this book and its content. We
are grateful to everyone who helped us discover, explain, and write down our
understanding of what matters in using statistics for data analysis. Although there
are many of you out there, we would like to thank a few in particular.
First, we are grateful for the support of the QUT’s School of Management,
School of Accountancy and Information Systems School, for supporting us in the
development and conduct of the Advanced Data Analysis workshop series; this
book would not have been possible without it. Special thanks go to Professor
Michael Rosemann and Professor Peter Green for their inspiring entrepreneurial
spirit, flexibility, and support. Second, we are grateful that so many of our
colleagues and students attended these workshops and discussed and challenged
our understanding of data analysis methods and the way we taught them.
Finally, the ones that contributed—or perhaps suffered—most are our lovely
wives, Laura, Claudia, and Laura. Thank you for your support, your patience, and
for sharing some of our headaches. You make our lives 89 % more enjoyable
(p < .001, [75–100]).

September 2016 Willem Mertens

Amedeo Pugliese
Jan Recker

vii
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction to the Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Navigating the World of Statistics—And This Book . . . . . . . . . . 3
1.3 What This Book Does Not Cover . . . . . . . . . . . . . . . . . . . . . . . 5
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Comparing Differences Across Groups . . . . . . . . . . . . . . . . . . . . . . 7
2.1 One or Two Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 More than Two Groups: One-Way ANOVA . . . . . . . . . . . . . . . . 10
2.3 More than Two Grouping Variables: Factorial ANOVA . . . . . . . 12
2.4 More than One Dependent Variable: Multivariate ANOVA . . . . 14
2.5 More Advanced Models: Covariance and Repeated Measures . . . 16
2.6 When to Use Group Comparisons . . . . . . . . . . . . . . . . . . . . . . . 17
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3 Assessing (Innocuous) Relationships . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1 What Are Regression Models? . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 When Do We Use Regression Models? . . . . . . . . . . . . . . . . . . . 24
3.3 How Do We Examine Regression Models? . . . . . . . . . . . . . . . . 26
3.4 How Do We Report Regression Analyses? . . . . . . . . . . . . . . . . . 31
3.5 What If. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4 Models with Latent Concepts and Multiple Relationships:
Structural Equation Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.1 What Are Structural Equation Models? . . . . . . . . . . . . . . . . . . . 37
4.2 When Do We Use Structural Equation Models? . . . . . . . . . . . . . 41
4.3 How Do We Examine Structural Equation Models? . . . . . . . . . . 43
4.4 How Do We Report Structural Equation Model Analyses? . . . . . 49
4.5 What If. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5 Nested Data and Multilevel Models: Hierarchical Linear Modeling . . . 61
5.1 What Are Hierarchical Linear Models? . . . . . . . . . . . . . . . . . . . 61
5.2 When Do We Use HLMs? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

ix
x Contents

5.3 How Do We Investigate HLMs? . . . . . . . . . . . . . . . . . . . . . . . . 64

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6 Analyzing Longitudinal and Panel Data . . . . . . . . . . . . . . . . . . . . . 73
6.1 What Are Longitudinal and Panel Data? . . . . . . . . . . . . . . . . . . 73
6.2 Clustering as a Way to Deal with Nestedness . . . . . . . . . . . . . . . 77
6.3 Which Models Can We Use to Analyze Longitudinal Data? . . . . 81
6.4 Estimating and Reporting Fixed-Effects and Random-Effects
Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.5 When to Use OLS, Fixed-Effects, and Random-Effects Models . . . 92
6.6 Final Remarks, Suggestions and Your Best Tool: Thinking! . . . . 96
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7 Causality: Endogeneity Biases and Possible Remedies . . . . . . . . . . 99
7.1 Your Research Question Is Causal: What Does that Mean? . . . . . 100
7.2 Self-Selection and Endogeneity . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.3 Specifying OLS Models to Minimize Endogeneity Concerns . . . 107
7.4 More Complex Designs to Support Causal Claims . . . . . . . . . . . 111
7.5 Some Caveats and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . 132
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
8 How to Start Analyzing, Test Assumptions and Deal
with that Pesky p-Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
8.1 Structuring, Cleaning, and Summarizing Data . . . . . . . . . . . . . . 136
8.1.1 Structuring Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
8.1.2 Cleaning Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8.1.3 Exploring Data: Summary Statistics and Visualization . . . 142
8.2 Testing Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
8.2.1 Independence of Observations . . . . . . . . . . . . . . . . . . . . 143
8.2.2 Normality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
8.2.3 Homogeneity of Variance and Homoscedasticity . . . . . . . 148
8.2.4 Linearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
8.2.5 What if Assumptions Are Violated? . . . . . . . . . . . . . . . . 150
8.3 Mindfully Interpreting Statistics: The Case of the p-Value . . . . . 152
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
9 Keeping Track and Staying Sane . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Introduction
1

Data, data, data. More data is available to us now than ever before. As work and
private activities are increasingly facilitated and enacted by our digital devices, we
leave traces that can be picked up and analyzed anytime and anywhere. Data
collected through surveys, archives, and experiments also remain relevant, as
digital traces do not necessarily reflect perceptions, attitudes, and intentions.
What also has not changed is that data is meaningless until it is analyzed. That is
what this book is about: analyzing data. More precisely: analyzing quantitative
data. Numbers.
Data analysis is an iterative process of manipulating and interpreting numbers to
extract meaning from them—answer research questions, test hypotheses, or explore
meanings that can be derived inductively from the data. This exploration is the first
step of any data analysis: we run a few basic manipulations and tests to summarize
the data in meaningful statistics, such as means and standard deviations; we
visualize the data; and we try to improve our understanding of the information in
the data.
Of course, before you can start analyzing, you need to obtain data and have a
rough idea of the meaning you want to extract through analysis. Therefore, every
chapter briefly discusses when and why you may want to use the methods discussed
in that chapter, including the type of questions typically answered and the type of
data analyzed using that method. We hope that this approach will help you
understand how theory, research designs, research questions, data, and analysis
depend on one another. The credibility of data is derived from the research design;
and the credibility of data analysis is derived from its grounding in theory.
Before we start discussing data analysis methods, we want to summarize the key
concepts used in this book and give you a roadmap for using the book and choosing
the right analysis method.

# Springer International Publishing Switzerland 2017 1

W. Mertens et al., Quantitative Data Analysis, DOI 10.1007/978-3-319-42700-3_1
2 1 Introduction

1.1 Introduction to the Basics

Before we dive into the wonderful world of applied empirical research and quanti-
tative data analysis, there are a few basic words and rules we will summarize
quickly. Everyone reading this book may already have this knowledge, but it
never hurts to make sure we speak the same language.
Let’s start with the very basics:
A case is a person, team, organization, IT system, or any subject or object that is
the primary unit of analysis. What constitutes a case is not given but an important
decision that we discuss in detail in Chap. 8 of this book, when we talk about
structuring data. Many other terms can be used to refer to a case, such as a unit, a
subject, a respondent, or any relevant unit of analysis (e.g., store, board member).
A variable is an attribute of a case that can change into any one of a set of values.
It is called a variable because the values change—the meaning of the variable varies
(hence the name). One of the variables related to humans is age because our age
changes throughout our lives and because one human’s age is not the same as
another’s. That makes age different from, say, the speed of light, which Einstein
showed us to be a constant—it never changes.
There are several types of variables: Categorical variables do not make
assumptions about the distance between the values that the variable may assume,
so the values can be nominal (e.g., someone’s role as “librarian” or “scientist”),
dichotomous (the choice between two values, e.g., gender), or ordinal (e.g., a
ranking of intelligence, from normal to literate to smart). Continuous variables
can take on any value, although there is a difference between truly continuous
variables (like time and speed) and discrete variables that jump between whole
values without decimals (such as age when measured in years). This difference is
important because many sciences (especially the social sciences) measure
responses on a quasi-continuous scale, such as a 5-point Likert-type response
scale in which the respondent chooses among, for example, “strongly disagree,”
“disagree,” “neutral,” “agree,” and “strongly agree.” These sciences treat such
choices as continuous variables, and we do the same in this book.
Most research in social sciences is conducted based on samples, not the whole
population. A sample is that part of a population that is subjected to observation
(i.e., measurement), which we analyze to draw inferences (i.e., draw conclusions
based on observation) about the population. For example, if we want to say
something about the IT-savviness of Baby Boomers (the population), we could
find a representative sample of, say, 250 60-year-olds, let them fill out a survey or
complete a number of tasks using digital devices while measuring their efficiency
and effectiveness, and use the data collected from that particular sample to draw
conclusions about Baby Boomers in general. We often call the analysis tools we use
to that end inferential statistics, because our aim is to infer (i.e., conclude) some-
thing about the population from the sample data we have. A representative sample
means that the characteristics of the sample reflect the characteristics of the larger
population. For example, a representative sample of Baby Boomers would not be
made up only of men, carpenters, or gadget freaks.
1.2 Navigating the World of Statistics—And This Book 3

Data refers to any set of values that measure a variable or the combination of all
variables for all cases in your study. Sometimes data is collected at several levels.
For example, we may measure certain attributes of individuals (level 1) or certain
attributes of the organizational department to which these individuals belong (level
2). Alternatively, we may measure the same attributes at several points in time.
Both scenarios lead to data that is nested—that is, the nested variables can have
multiple values for a single value of the variables within which they are nested.
(Hmmm. . .that was clear as mud, wasn’t it?) For example, multiple individuals are
nested within one department, so the variable “staff number” is be nested within
“department,” and one department will have multiple values for “staff number.”
Got it?
Data collected from cases that are part of your sample are used to test
hypotheses, which are theoretical statements or predictions that you make about
relationships between variables that you expect to occur in the population. For
example, a hypothesis about the relationship between the variables “tech savviness”
and “age” might be that tech savviness decreases as age increases.
Test statistics help you determine whether the hypotheses are valid by assessing
the observed relationships between variables in your sample in light of estimated
summary statistics, such as means and standard deviations. Means are the expected
values of the variables, while standard deviations are the average distance between
the other values you find and that mean. The p-value that is reported as part of the
test statistics usually refers to the probability that a test statistic’s value would be as
high or low if it were only chance that was at play. To conclude the significance of
an effect we compare the p-value to a predefined alpha value, which is the point at
which p is considered too small to be determined only by chance (usually 0.05 or
0.01). However, p-values and their corresponding statistics are meaningless without
context, as they are but one of many elements that make results relevant and
credible. Therefore, they should be reported only along with extensive information
about data distributions, effect sizes and the approaches that were used to collect
and analyze the data.
One of the many reasons that such is the case is that relying solely on p-values
exposes us to the risk of making Type I and Type II errors. A Type I error occurs
when you conclude that there is a significant effect when there is none (e.g., a doctor
finds that a male patient is pregnant); a Type II error occurs when you find no result
when there is one (e.g., a doctor finds that a woman in labor is not pregnant).
If that high-level review leaves you baffled, don’t give up! We will go into more
detail about each of these terms in the following chapters.

1.2 Navigating the World of Statistics—And This Book

This book is largely about finding the right method for your data and research
questions. Figure 1.1 presents a decision framework that is all you will require
to achieve just that. So why do you need to read the rest of the book? Proper
use of the framework requires a basic knowledge of its components. That is why
4 1 Introduction

Do you
Do you understand
have data? Yes No
your data?

No

Do you Chapter 6:
study change Yes Longitudinal and
over time? Panel Data Analyses

No

Do you Chapter 5: Did you Chapter 8:

study relations Yes Hierarchical Linear check relevant No Testing
that cross Models (HLM) assumptions? Assumptions
levels?

No

Do you Chapter 4:
study latent Yes Structural Equation
concepts? Modeling (SEM)

No

Do you
study two Chapter 3: Comparing Do you Chapter 7:
or more Yes Differences Across worry about Yes Endogeneity
groups? Groups causality?

No

Do you
study how Chapter 2:
two or more Yes Assessing Innocuous
variables Relationships
relate?

No

This book is
not for you!

Fig. 1.1 How to use this book

we recommend using this book in an iterative way: use the decision tree to find the
method that seems appropriate, read about it, and then reassess your choice. We
hope that one such iteration will be enough to solve your most pressing questions,
but experience tells us that it will not.
1.3 What This Book Does Not Cover 5

Our recommendation to use the decision tree to determine which chapters to read
means that each of the chapters is written so it can be read independently of the
others. However, keep in mind that the chapters become increasingly complex,
reflecting the increasing complexity of the methods discussed throughout the book.
As Figure 1.1 shows, some of the fundamental challenges in quantitative data
analysis are discussed only toward the end of this book: Can I prove causality?
What exactly is contained in my data? What assumptions do I need to be aware of?
How do I deal with the ongoing discussion on the use of p-values? These topics can
be complex and tedious, so if you are very new to statistics, we recommend that you
browse through the book from start to finish (or answer “no” to every question you
do not understand). If you follow this approach and make it to the last chapter of the
book, we will reward you with some practical hints and tips on keeping track of
your analyses and keeping your mental health as you struggle with mastering
complex statistics. Based on how complex your questions and data are, the decision
tree takes you backward, step by step, from fairly complex analyses (evaluations
involving time, panels, levels, and latent concepts) to fairly simple analyses
(differences between groups and relationships between variables).

1.3 What This Book Does Not Cover

We should probably warn you about all the wonderful things you will not find in
this book. We start by reiterating that this is not a book about statistics; it’s a book
about using statistics. We use few equations and then only when we think they can
help you understand a method. This is not a manual either: for example, we do not
provide step-by-step recipes for the perfect regression using SPSS. Although we
provide guidance on how to apply methods, we will refer you to other resources for
more detailed help.
The book will also not help you design the perfect quantitative study or data
collection procedure. Although we provide some advice on both research design
and data collection, this advice is inspired by and structured around analysis. More
extensive, start-to-end advice on those topics can be found in Recker [4].
Another thing this book does not provide is an exhaustive overview of methods,
as we discuss only a selection of methods that are popular in the Accounting and
Information Systems research fields. Some analyses, such as discriminant analyses,
are too fancy for us to cover, while others, like Bayesian updating, are not very
popular in our fields. We also rely heavily on parametric and linear analyses and
steer mostly clear of non-parametric and nonlinear analyses.
Finally, in most of what we discuss we concentrate on simple and straightfor-
ward scenarios without discussing every little option, variant, or exception. Once
you understand the basis of a method, variants and exceptions will not be much of a
challenge.
And now on to the rest of the book!
6 1 Introduction

References
1. Crawley MJ (2013) The R book, 2nd edn. Wiley, West Sussex
2. Field AP (2013) Discovering statistics using IBM SPSS statistics: and sex and drugs and rock
‘n’ roll, 4th edn. Sage, London
3. Stevens JP (2009) Applied multivariate statistics for the social sciences. Taylor and Francis
Group, LLC, London
4. Recker J (2012) Scientific research in information systems: A beginner’s guide. Springer,
Heidelberg
Comparing Differences Across Groups
2

Imagine you want to find out whether people who read statistics books are better at
analyzing data than those who do not. You could study this question in at least three
ways: (1) You set up an experiment with two groups, one of which you make read a
statistics book (a real cruelty), and then you make both groups analyze the same
data. (2) You run an experiment with one group only, test their analysis skills, make
them read a statistics book, and then test their skills again. (3) You find some people
who read statistics books and other people who do not and compare their analysis
skills by, for example, studying the number of their quantitative research
publications or their grades in statistics classes. All three ways would end up with
one variable that tells you whether a person reads statistics books or not—a
dichotomous variable that defines group membership—and one continuous variable
that summarizes people’s analysis skills (or statistics performance). Answering
your research question would require you to evaluate whether the analysis skills
of the group that read the book are better than those of the other group. This form of
group comparisons—comparing one variable score between two groups—is the
simplest. This chapter starts from this simple example and adds complexity by
adding more groups and variables of interest.

2.1 One or Two Groups

The research question concerning whether people who read statistics books are
better at analyzing data than those who do not, which requires comparing two
groups, is typically answered by using a t-test. A t-test takes the average score of
one group as a reference point and determines whether a second group’s average
score differs from that of the first one by estimating the distance between the two
means and comparing that distance to the variance in the data. Variance, a measure
of how spread out data points are, is calculated by summing up the squared
difference between the mean and every raw score (i.e., my skills minus the average,
your skills minus the average, then multiplying each of these differences by itself,

# Springer International Publishing Switzerland 2017 7

W. Mertens et al., Quantitative Data Analysis, DOI 10.1007/978-3-319-42700-3_2
8 2 Comparing Differences Across Groups

which helps to get rid of “negative distances” when the score is smaller than the
mean. This process results in one measure of the average squared distance of data
points to the mean). A t-test calculates the difference between two group means and
compares it to the average distance of all data points to the mean.1
There are three kinds of t-tests, each fitting a certain research design. The first
kind is the independent-samples t-test, which you would use for the first (1) and
third (3) study outlined in the first paragraph of this chapter. The independent-
samples t-test compares the average scores of two independent samples, that is,
samples whose scores were derived from separate groups of people who were not
influenced by each other.2
A different situation exists in option (2), so it uses a second kind of t-test, the
paired-samples t-test. In the option (2) scenario, there is only one group, but the
group’s skills are tested twice, once before and once after reading a statistics book.
These two scores are not independent, as the later score depends to some extent on
the earlier score; both scores are for the same group at different points in time.
The third kind of t-test is the one-sample t-test. Although it still compares two
means, is does so based on data from one sample only. Imagine, for example, that
we are interested in the effect of one particular statistics book only, and that this
book has not been published yet. Imagine also that there is one widely accepted test
to measure analysis skills for which an overall average (a population average) is
available.3 In this case, we would let one group of people read the book, test their
skills using the standard skills test, and compare their average score with the
publicly available average score of all other people who have taken the skills test.
We are still comparing two groups, but we collect data from only one sample.

How to Analyze and Report Simple Group Comparisons

Imagine we collected the data shown in Table 2.1. A good first step is always to
calculate some descriptive statistics. Say that these descriptive statistics tell us that
the mean score of people who read statistics is higher (M1 ¼ 67.25) than that of
people who do not (M2 ¼ 55.75) and that the standard deviation (the square root of
the variance) for both groups is similar at first look (SD1 ¼ 19.47; SD2 ¼ 18.51).
As is the case with most of the tests discussed in this book, the significance of the
difference between group averages depends in part on the variance within both
groups. If these variances are large, a small difference between the two groups is
likely to be insignificant (in other words, it would not be unexpected to find a larger
difference). However, if the variances are small (meaning that the individuals in the
group have similar analytical skills), even a small difference in the average

1
In fact, depending on the kind of t-test, the difference between the means is usually compared
only to the variance in part of the collected data.
2
For more information on independence of samples, please refer to Sect. 8.2.
3
A well-known example of a group of tests for which such reference scores are available is IQ
tests. Most common IQ tests have a population average of about 100, so it is easy to evaluate
individuals’ or groups’ scores against that average. A one-sample t-test would allow you to
compare a group average against the population average score of 100.
2.1 One or Two Groups 9

Table 2.1 Analytic skills Person ID Read stats books? Analytic skills
test scores—example
1 No 32
2 Yes 61
3 Yes 93
4 Yes 62
5 No 65
6 No 62
7 No 73
8 Yes 87
9 Yes 39
10 No 86
11 Yes 81
... ... ...

analytical skills between the two groups—one that does and one that does not read
statistics books—could be considered significant.
Because variance plays such an important role, you have to test whether the
variances in the two groups are close enough to consider the two sets of data
homogenous. Homogeneity of variances is one of three assumptions that must be
checked. (See Sect. 8.2 to learn what assumptions are and how to check them.) In
normal circumstances, only the variance of one group is taken into account when
calculating the differences between groups. However, if the homogeneity of
variances is violated, the one group is not a good representation of the variance in
all of the data. In that case, we would have to specify that the variances are not
homogenous when calculating the t-test so both variances are taken into account.
The third assumption, alongside independence of samples and equality of
variances, is that scores are normally distributed. The t-test is pretty robust against
minor violations of this last assumption, but as always it is best to check.
Once these assumptions have been checked, it is time to specify your
hypotheses. The null hypothesis (H0) specifies the reference point—the absence
of an effect, while the alternative hypothesis (H1) postulates that there is a differ-
ence between means.

H0: The analytic skills of people who read statistics books are equal to those of
people who do not read statistics books.
H1: The analytic skills of people who read statistics books are superior to those of
people who do not read statistics books.

The alternative hypothesis (H1) shown here is an example of a hypothesis that

should be tested with a one-sided t-test, as we are interested only in whether there is
a positive difference between people who do read statistics books and those who do
not. If we did not care whether their skills are better or worse but only in whether
there is a difference at all, we could use a two-sided t-test.
10 2 Comparing Differences Across Groups

Most statistical software packages allow t-statistics to be calculated in a few

straightforward steps. Typically, you specify whether you are interested in a
one-sided or two-sided t-test, the α (the threshold at which p is significant—
typically 0.05 or 0.01), the reference group, and the test group, and you run the
test. Executing these steps based on our fictional data would tell you that people
who read statistics books (Mdo ¼ 67.25) do not have significantly better analytic
skills than people who do not (Mdo not ¼ 55.75; t ¼ 1.48, p ¼ 0.152). This example
also shows how t-tests are typically reported on: You report descriptive statistics
(at least the number of participants, the means, and the standard deviations), specify
whether it is a one- or two-sided test, present your hypothesis, and report the value
for the t-statistic and the level of p at which it was considered significant. If the
value was significant, it is customary to report the alpha at which it was significant
(e.g. p < 0.05), rather than the exact p-value. These guidelines also apply to group
comparisons with more than two groups and more than two variables, although a
larger number of statistics will be reported in such cases.

2.2 More than Two Groups: One-Way ANOVA

We now know that people who read a statistics book do not have significantly better
analytic skills than people who do not, but reading one statistics book may simply
not be enough. Would reading multiple books help? To test whether such is the
case, we could compare the analytic skills of a group that has never read a statistics
book, a group that has read one book, and a group that has read two or more books.
A t-test is not sufficient to evaluate the differences in skill among three groups, so
we use a one-way ANalysis Of VAriance (ANOVA).
Like t-tests, ANOVA models test differences among groups. A disadvantage is
that ANOVA tests are always two-sided, while an advantage is that they allow the
differences between more than two groups to be tested. As the name suggests, one-
way ANOVA models evaluate the differences between groups based on one group-
ing variable and an analysis of the mean and the variance. In doing so, it relies on
the F-statistic, which compares the amount of variance that can be explained by
group membership to the amount of variance that cannot be explained by the group.
Thus, if the F-statistic is low (indicated by a high p-value), splitting the sample into
groups does not help to clarify the data much. In our example, a low F-statistic
would mean that every group had comparable or seemingly random means and
variances, so reading statistics books is not related to people’s analytic skills.
Because both the mean and the variance are taken into account, ANOVA is pretty
robust against violations of the assumptions of homoscedasticity, independence of
observations, and normality of observations within the groups [1]. However,
regardless of its robustness, it’s always necessary to check (and it’s always
recommended that one correct) (see Sect. 8.2) for violations of the assumptions
that underlie the tests.
Similar to t-tests, the null hypothesis in ANOVA states that there are no
differences between groups; rejecting that null hypothesis in our example means
2.2 More than Two Groups: One-Way ANOVA 11

Table 2.2 Analytic skills Person ID # of statistics books read Analytic skills
test scores—example with
1 2 76
three groups
2 1 83
3 1 56
4 0 51
... ... ...

that the number of statistics books someone reads does affect his or her analytic
skills. Note that the causality implied in this result is supported by the experimental
design only; if the design does not involve an experiment with random assignment
of participants to conditions, ANOVA does not allow conclusions to be drawn
about the direction of relationships. In that case, we would conclude that the
number of statistics books someone reads is related to his or her analytical skills,
rather than that reading statistics books causes better analytical skills. We don’t
know what causes what.

How to Run and Report on a One-Way ANOVA Table 2.2 shows an excerpt of
sample data for our example. We now have three groups of people divided
according to the number of statistics books they have read—none, one, or more
than one. Running a one-way ANOVA test based on this data is, again, pretty
straightforward. The variable that defines group membership—in this case, the
number of books read—is included as the “factor” or the independent variable.
The dependent variable in our example is the participants’ analytic skills. Most
statistical packages will return an output that includes some version of within-group
and between-group sums of squares, which represent the squared average distances
to the relevant mean and are used to calculate the F-statistic. Refer to Field ([2])4 if
you want to know the ins and outs of ANOVA, its sums of squares, and the exact
calculation of the F-statistic.
If we had a sample of 101 students, the result of the ANOVA test could be
reported as follows: “There was a significant relationship between the number of
books students read and their analytic skills [F(2,98) ¼ 6.05, p < 0.01].” This result
suggests that there is a difference between at least two of the three groups. The
2 and the 98 are the degrees of freedom5 for the F-test. The first number, the 2, is
calculated based on the number of groups (k) that are compared, minus one (k  1);
the second number, the 98, is the number of people in the sample (n), minus the
number of groups compared (n  k). Statistical computer programs will always
report these numbers alongside other results. Because the degrees of freedom affect

4
Field [2] is an excellent source of detailed guidelines and probably the most entertaining statistics
resource ever written.
5
Although a discussion of degrees of freedom is outside of the scope of this book, the easiest way
to explain degrees of freedom is as the number of pieces of information that are taken into account
when estimating a statistic or a (set of) parameter(s). Therefore, degrees of freedom are usually
calculated based on the sample size and the number of parameters that are estimated.
Another random document with
no related content on Scribd:
 “For the present,” wrote the secretary to Rodgers,[326] “it is
desirable that with the force under your command you remain in such
position as to enable you most conveniently to receive further more
extensive and more particular orders, which will be conveyed to you
through New York. But as it is understood that there are one or more
British cruisers on the coast in the vicinity of Sandy Hook, you are at
your discretion free to strike them, returning immediately after into
port. You are free to capture or destroy them.”
These orders reached New York June 21. Rodgers in his fine
frigate the “President,” with the “Hornet,” 18, was eager to sail. The
hope of capturing the “Belvidera,” which had long been an
intolerable annoyance to New York commerce, was strong both in
the Navy Department and in the navy; but the chance of obtaining
prize money from the British West India convoy, just then passing
eastward only a few days’ sail from the coast, added greatly to the
commodore’s impatience.[327] Decatur’s squadron arrived off Sandy
Hook June 19. June 21, within an hour after receiving the secretary’s
orders of June 18, the whole fleet, including two forty-four and one
thirty-eight-gun frigates, with the “Hornet” and the “Argus,” stood
out to sea.
The secretary might have spared himself the trouble of giving
further orders, for many a week passed before Rodgers and Decatur
bethought themselves of his injunction to return immediately into
port after striking the “Belvidera.” They struck the “Belvidera” within
forty-eight hours, and lost her; partly on account of the bursting of
one of the “President’s” main-deck guns, which blew up the
forecastle deck, killing or wounding sixteen men, including
Commodore Rodgers himself, whose leg was broken; partly, and
according to the British account chiefly, on account of stopping to
fire at all, when Rodgers should have run alongside, and in that case
could not have failed to capture his enemy. Whatever was the
reason, the “Belvidera” escaped; and Rodgers and Decatur, instead
of returning immediately into port as they had been ordered, turned
in pursuit of the British West India convoy, and hung doggedly to the
chase without catching sight of their game, until after three weeks’
pursuit they found themselves within a day’s sail of the British
Channel and the convoy safe in British waters.
This beginning of the naval war was discouraging. The American
ships should not have sailed in a squadron, and only their good luck
saved them from disaster. Rodgers and Decatur showed no regard to
the wishes of the Government, although had they met with
misfortune, the navy would have lost its last hope. Yet if the two
commodores had obeyed the secretary’s commands their cruise
would probably have been in the highest degree disastrous. The
Government’s true intentions have been a matter of much dispute;
but beyond a doubt the President and a majority of his advisers
inclined to keep the navy within reach at first,—to use them for the
protection of commerce, to drive away the British blockaders; and
aware that the British naval force would soon be greatly increased,
and that the American navy must be blockaded in port, the
Government expected in the end to use the frigates as harbor
defences rather than send them to certain destruction.
With these ideas in his mind Secretary Hamilton, in his orders of
June 18, told Rodgers and Decatur that “more extensive” orders
should be sent to them on their return to New York. A day or two
afterward Secretary Gallatin complained to the President that these
orders had not been sent.
“I believe the weekly arrivals from foreign ports,” said Gallatin,
[328] “will for the coming four weeks average from one to one-and-a-
half million dollars a week. To protect these and our coasting vessels,
while the British have still an inferior force on our coasts, appears to
me of primary importance. I think that orders to that effect, ordering
them to cruise accordingly, ought to have been sent yesterday, and
that at all events not one day longer ought to be lost.”
June 22 the orders were sent according to Gallatin’s wish. They
directed Rodgers with his part of the squadron to cruise from the
Chesapeake eastwardly, and Decatur with his ships to cruise from
New York southwardly, so as to cross and support each other and
protect with their united force the merchantmen and coasters
entering New York harbor, the Delaware, and the Chesapeake.
Rodgers and Decatur were then beginning their private cruise across
the ocean, and never received these orders until the commerce they
were to protect either reached port in safety or fell into British
hands.
Probably this miscarriage was fortunate, for not long after
Rodgers and Decatur passed the Banks the British Vice-Admiral
Sawyer sent from Halifax a squadron to prevent the American navy
from doing what Secretary Hamilton had just ordered to be done.
July 5 Captain Broke, with his own frigate the “Shannon,” 38, the
“Belvidera,” 36, the “Africa,” 64, and “Æolus,” 32, put to sea from
Halifax and was joined, July 9, off Nantucket by the “Guerriere,” 38.
Against such a force Rodgers and Decatur, even if together, would
have risked total destruction, while a success would have cost more
than it was worth. The Americans had nothing to gain and
everything to lose by fighting in line-of-battle.
As Broke’s squadron swept along the coast it seized whatever it
met, and July 16 caught one of President Jefferson’s 16-gun brigs,
the “Nautilus.” The next day it came on a richer prize. The American
navy seemed ready to outstrip the army in the race for disaster. The
“Constitution,” the best frigate in the United States service, sailed
into the midst of Broke’s five ships. Captain Isaac Hull, in command
of the “Constitution,” had been detained at Annapolis shipping a new
crew, until July 5,[329]—the day when Broke’s squadron left Halifax;
—then the ship got under way and stood down Chesapeake Bay on
her voyage to New York. The wind was ahead and very light. Not till
July 10 did the ship anchor off Cape Henry lighthouse,[330] and not
till sunrise of July 12 did she stand to the eastward and northward.
Light head-winds and a strong current delayed her progress till July
17, when at two o’clock in the afternoon, off Barnegat on the New
Jersey coast, the lookout at the masthead discovered four sails to
the northward, and two hours later a fifth sail to the northeast. Hull
took them for Rodgers’s squadron. The wind was light, and Hull
being to windward determined to speak the nearest vessel, the last
to come in sight. The afternoon passed without bringing the ships
together, and at ten in the evening, finding that the nearest ship
could not answer the night signal, Hull decided to lose no time in
escaping.
Then followed one of the most exciting and sustained chases
recorded in naval history. At daybreak the next morning one British
frigate was astern within five or six miles, two more were to
leeward, and the rest of the fleet some ten miles astern, all making
chase. Hull put out his boats to tow the “Constitution;” Broke
summoned the boats of his squadron to tow the “Shannon.” Hull
then bent all his spare rope to the cables, dropped a small anchor
half a mile ahead, in twenty-six fathom water, and warped his ship
along. Broke quickly imitated the device, and slowly gained on the
chase. The “Guerriere” crept so near Hull’s lee-beam as to open fire,
but her shot fell short. Fortunately the wind, though slight, favored
Hull. All night the British and American crews toiled on, and when
morning came the “Belvidera,” proving to be the best sailer, got in
advance of her consorts, working two kedge-anchors, until at two
o’clock in the afternoon she tried in her turn to reach the
“Constitution” with her bow guns, but in vain. Hull expected capture,
but the “Belvidera” could not approach nearer without bringing her
boats under the “Constitution’s” stern guns; and the wearied crews
toiled on, towing and kedging, the ships barely out of gunshot, till
another morning came. The breeze, though still light, then allowed
Hull to take in his boats, the “Belvidera” being two and a half miles
in his wake, the “Shannon” three and a half miles on his lee, and the
three other frigates well to leeward. The wind freshened, and the
“Constitution” drew ahead, until toward seven o’clock in the evening
of July 19 a heavy rain-squall struck the ship, and by taking skilful
advantage of it Hull left the “Belvidera” and “Shannon” far astern;
yet until eight o’clock the next morning they were still in sight
keeping up the chase.
Perhaps nothing during the war tested American seamanship
more thoroughly than these three days of combined skill and
endurance in the face of an irresistible enemy. The result showed
that Hull and the “Constitution” had nothing to fear in these
respects. There remained the question whether the superiority
extended to his guns; and such was the contempt of British naval
officers for American ships, that with this experience before their
eyes they still believed one of their 38-gun frigates to be more than
a match for an American forty-four, although the American, besides
the heavier armament, had proved his capacity to out-sail and out-
manœuvre the Englishman. Both parties became more eager than
ever for the test. For once, even the Federalists of New England felt
their blood stir; for their own President and their own votes had
called these frigates into existence, and a victory won by the
“Constitution,” which had been built by their hands, was in their eyes
a greater victory over their political opponents than over the British.
With no halfhearted spirit, the sea-going Bostonians showered well-
weighed praises on Hull when his ship entered Boston harbor, July
26, after its narrow escape; and when he sailed again, New England
waited with keen interest to learn his fate.
Hull could not expect to keep command of the “Constitution.”
Bainbridge was much his senior, and had the right to a preference in
active service. Bainbridge then held and was ordered to retain
command of the “Constellation,” fitting out at the Washington Navy
Yard; but Secretary Hamilton, July 28, ordered him to take command
also of the “Constitution” on her arrival in port. Doubtless Hull
expected this change, and probably the expectation induced him to
risk a dangerous experiment; for without bringing his ship to the
Charlestown Navy Yard, but remaining in the outer harbor, after
obtaining such supplies as he needed, August 2, he set sail without
orders, and stood to the eastward. Having reached Cape Race
without meeting an enemy he turned southward, until on the night
of August 18 he spoke a privateer, which told him of a British frigate
near at hand. Following the privateersman’s directions the
“Constitution” the next day, August 19, at two o’clock in the
afternoon, latitude 41° 42´, longitude 55° 48´, sighted the
“Guerriere.”
The meeting was welcome on both sides. Only three days before,
Captain Dacres had entered on the log of a merchantman a
challenge to any American frigate to meet him off Sandy Hook. Not
only had the “Guerriere” for a long time been extremely offensive to
every seafaring American, but the mistake which caused the “Little
Belt” to suffer so seriously for the misfortune of being taken for the
“Guerriere” had caused a corresponding feeling of anger in the
officers of the British frigate. The meeting of August 19 had the
character of a preconcerted duel.
The wind was blowing fresh from the northwest, with the sea
running high. Dacres backed his main-top-sail and waited. Hull
shortened sail and ran down before the wind. For about an hour the
two ships wore and wore again, trying to get advantage of position;
until at last, a few minutes before six o’clock, they came together
side by side, within pistol-shot, the wind almost astern, and running
before it they pounded each other with all their strength. As rapidly
as the guns could be worked, the “Constitution” poured in broadside
after broadside, double-shotted with round and grape,—and, without
exaggeration, the echo of these guns startled the world. “In less
than thirty minutes from the time we got alongside of the enemy,”
reported Hull,[331] “she was left without a spar standing, and the
hull cut to pieces in such a manner as to make it difficult to keep her
above water.”
That Dacres should have been defeated was not surprising; that
he should have expected to win was an example of British arrogance
that explained and excused the war. The length of the “Constitution”
was 173 feet; that of the “Guerriere” was 156 feet; the extreme
breadth of the “Constitution” was 44 feet; that of the “Guerriere”
was 40 feet, or within a few inches in both cases. The “Constitution”
carried thirty-two long 24-pounders, the “Guerriere” thirty long 18-
pounders and two long 12-pounders; the “Constitution” carried
twenty 32-pound carronades, the “Guerriere” sixteen. In every
respect, and in proportion of ten to seven, the “Constitution” was
the better ship; her crew was more numerous in proportion of ten to
six. Dacres knew this very nearly as well as it was known to Hull, yet
he sought a duel. What he did not know was that in a still greater
proportion the American officers and crew were better and more
intelligent seamen than the British, and that their passionate wish to
repay old scores gave them extraordinary energy. So much greater
was the moral superiority than the physical, that while the
“Guerriere’s” force counted as seven against ten, her losses counted
as though her force were only two against ten.
Dacres’ error cost him dear, for among the “Guerriere’s” crew of
two hundred and seventy-two, seventy-nine were killed or wounded;
and the ship was injured beyond saving before Dacres realized his
mistake, although he needed only thirty minutes of close fighting for
the purpose. He never fully understood the causes of his defeat, and
never excused it by pleading, as he might have done, the great
superiority of his enemy.[332]
Hull took his prisoners on board the “Constitution],” and after
blowing up the “Guerriere” sailed for Boston, where he arrived on
the morning of August 30. The Sunday silence of the Puritan city
broke into excitement as the news passed through the quiet streets
that the “Constitution” was below, in the outer harbor, with Dacres
and his crew prisoners on board. No experience of history ever went
to the heart of New England more directly than this victory, so
peculiarly its own; but the delight was not confined to New England,
and extreme though it seemed it was still not extravagant, for
however small the affair might appear on the general scale of the
world’s battles, it raised the United States in one half hour to the
rank of a first-class Power in the world.
Hull’s victory was not only dramatic in itself, but was also
supremely fortunate in the moment it occurred. The “Boston Patriot”
of September 2, which announced the capture of the “Guerriere,”
announced in the next column that Rodgers and Decatur, with their
squadron, entered Boston harbor within four-and-twenty hours after
Hull’s arrival, returning empty-handed after more than two months
of futile cruising; while in still another column the same newspaper
announced “the melancholy intelligence of the surrender of General
Hull and his whole army to the British General Brock.” Isaac Hull was
nephew to the unhappy General, and perhaps the shattered hulk of
the “Guerriere,” which the nephew left at the bottom of the Atlantic
Ocean, eight hundred miles east of Boston, was worth for the
moment the whole province which the uncle had lost, eight hundred
miles to the westward; it was at least the only equivalent the people
could find, and they made the most of it. With the shock of new life,
they awoke to the consciousness that after all the peace teachings
of Pennsylvania and Virginia, the sneers of Federalists and
foreigners; after the disgrace of the “Chesapeake” and the surrender
of Detroit,—Americans could still fight. The public had been taught,
and had actually learned, to doubt its own physical courage; and the
reaction of delight in satisfying itself that it still possessed the
commonest and most brutal of human qualities was the natural
result of a system that ignored the possibility of war.
Hull’s famous victory taught the pleasures of war to a new
generation, which had hitherto been sedulously educated to think
only of its cost. The first taste of blood maddens; and hardly had the
“Constitution” reached port and told her story than the public
became eager for more. The old Jeffersonian jealousy of the navy
vanished in the flash of Hull’s first broadside. Nothing would satisfy
the craving of the popular appetite but more battles, more British
frigates, and more daring victories. Even the cautious Madison was
dragged by public excitement upon the element he most heartily
disliked.
The whole navy, was once more, September 1, safe in port,
except only the “Essex,” a frigate rated at thirty-two but carrying
forty-four guns, commanded by Captain David Porter. She left New
York, July 3, with orders,[333] dated June 24, to join Rodgers, or
failing this to cruise southwardly as far as St. Augustine. June 11 she
met a convoy of seven transports conveying a battalion of the First
Regiment, or Royal Scots, from the West Indies to reinforce Prevost
and Brock in Canada. Porter cut out one transport. With the aid of
another frigate he could have captured the whole, to the great
advantage of Dearborn’s military movements; but the British
commander managed his convoy so well that the battalion escaped,
and enabled Prevost to strengthen the force at Niagara which
threatened and defeated Van Rensselaer. August 13 the British 20-
gun sloop-of-war “Alert” came in sight, bore down within short
pistol-shot, and opened fire on the “Essex.” Absurd as the idea
seemed, the British captain behaved as though he hoped to capture
the American frigate, and not until Porter nearly sunk him with a
broadside did the Englishman strike his colors. After taking a number
of other prizes, but without further fighting, September 7 Porter
brought his ship back to the Delaware River.
The return of the “Essex” to port, September 7, brought all the
national vessels once more under the direct control of the
Department. Nearly every ship in the service was then at Boston.
The three forty-fours—the “Constitution,” “United States,” and
“President”—were all there; two of the thirty-eights—the “Congress”
and “Chesapeake”—were there, and the “Constellation” was at
Washington. The “Adams,” 28, was also at Washington; but the
“Hornet,” 18, and “Argus,” 16, were with Rodgers and Decatur at
Boston. The “Syren,” 16, was at New Orleans; the “Essex,” 32, and
the “Wasp,” 18, were in the Delaware.
Carried away by Hull’s victory, the Government could no longer
hesitate to give its naval officers the liberty of action they asked, and
which in spite of orders they had shown the intention to take. A new
arrangement was made. The vessels were to be divided into three
squadrons, each consisting of one forty-four, one light frigate, and
one sloop-of-war. Rodgers in the “President” was to command one
squadron, Bainbridge in the “Constitution” was to command another,
and Decatur in the “United States” was to take the third.[334] Their
sailing orders, dated October 2,[335] simply directed the three
commodores to proceed to sea: “You are to do your utmost to annoy
the enemy, to afford protection to our commerce, pursuing that
course which to your best judgment may under all circumstances
appear the best calculated to enable you to accomplish these objects
as far as may be in your power, returning into port as speedily as
circumstances will permit consistently with the great object in view.”
Before continuing the story of the frigates, the fate of the little
“Wasp” needs to be told. Her career was brief. The “Wasp,” a sloop-
of-war rated at eighteen guns, was one of President Jefferson’s
additions to the navy to supply the loss of the “Philadelphia;” she
was ship-rigged, and armed with two long 12-pounders and sixteen
32-pound carronades. She carried a crew of one hundred and thirty-
seven men, commanded by Captain Jacob Jones, a native of
Delaware, lieutenant in the “Philadelphia” when lost in the war with
Tripoli. The “Wasp” was attached to Rodgers’s squadron, and
received orders from the commodore to join him at sea. She sailed
from the Delaware October 13, and when about six hundred miles
east of Norfolk, October 17, she fell in with the British 18-gun brig
“Frolic,” convoying fourteen merchantmen to England. The two
vessels were equal in force, for the “Frolic’s” broadside threw a
weight of two hundred and seventy-four pounds, while that of the
“Wasp” threw some few pounds less; the “Frolic” measured, by
British report,[336] one hundred feet in length, the “Wasp” one
hundred and six; their breadth on deck was the same; and although
the “Wasp’s” crew exceeded that of her enemy, being one hundred
and thirty-five men against one hundred and ten, the British vessel
had all the men she needed, and suffered little from this inferiority.
The action began at half-past eleven in the morning, the two sloops
running parallel, about sixty yards apart, in a very heavy sea, which
caused both to pitch and roll so that marksmanship had the most
decisive share in victory. The muzzles of the guns went under water,
and clouds of spray dashed over the crews, while the two vessels
ran side by side for the first fifteen minutes. The British fire cut the
“Wasp’s” rigging, while the American guns played havoc with the
“Frolic’s” hull and lower masts. The vessels approached each other
so closely that the rammers of the guns struck the enemy’s side, and
at last they fell foul,—the “Wasp” almost squarely across the
“Frolic’s” bow. In the heavy sea boarding was difficult; but as soon
as the “Wasp’s” crew could clamber down the “Frolic’s” bowsprit,
they found on the deck the British captain and lieutenant, both
severely wounded, and one brave sailor at the wheel. Not twenty of
the British crew were left unhurt, and these had gone below to
escape the American musketry. The “Wasp” had only ten men killed
and wounded. The battle lasted forty-three minutes.
If the American people had acquired a taste for blood, the battle
of the “Wasp” and “Frolic” gratified it, for the British sloop was
desperately defended, and the battle, won by the better
marksmanship of the Americans, was unusually bloody. Captain
Jones lost the full satisfaction of his victory, for a few hours
afterward the “Poictiers,” a British seventy-four, came upon the two
disabled combatants and carried both into Bermuda; but the
American people would have been glad to part with their whole navy
on such terms, and the fight between the “Wasp” and the “Frolic”
roused popular enthusiasm to a point where no honors seemed to
satisfy their gratitude to Captain Jones and his crew.
The “Wasp’s” brilliant career closed within a week from the day
she left the Delaware. A week afterward another of these ship-duels
occurred, which made a still deeper impression. Rodgers and
Decatur sailed from Boston October 8, with the “President,” the
“United States,” “Congress,” and “Argus,” leaving the “Constitution,”
“Chesapeake,” and “Hornet” in port. Rodgers in the “President,” with
the “Congress,” cruised far and wide, but could find no enemy to
fight, and after making prize of a few merchantmen returned to
Boston, December 31. The “Argus” also made some valuable prizes,
but was chased by a British squadron, and only by excellent
management escaped capture, returning Jan. 3, 1813, to New York.
Decatur in the “United States,” separating from the squadron
October 12, sailed eastward to the neighborhood of the Azores,
until, October 25, he sighted a sail to windward. The stranger made
chase. The wind was fresh from south-southeast, with a heavy sea.
Decatur stood toward his enemy, who presently came about, abreast
of the “United States” but beyond gunshot, and both ships being
then on the same tack approached each other until the action began
at long range. The British ship was the 38-gun frigate “Macedonian”
commanded by Captain Carden, and about the same force as the
“Guerriere.” At first the “United States” used only her long 24-
pounders, of which she carried fifteen on her broadside, while the
“Macedonian” worked a broadside of fourteen long 18-pounders. So
unequal a contest could not continue. Not only was the American
metal heavier, but the American fire was quicker and better directed
than that of the Englishman; so that Carden, after a few minutes of
this experience, bore down to close. His manœuvre made matters
worse. The carronades of the “United States” came into play; the
“Macedonian’s” mizzen-mast fell, her fore and main top-mast were
shot away, and her main-yard; almost all her rigging was cut to
pieces, and most of the guns on her engaged side were dismounted.
She dropped gradually to leeward, and Decatur, tacking and coming
up under his enemy’s stern, hailed, and received her surrender.
The British ship had no right to expect a victory, for the disparity
of force was even greater than between the “Constitution” and
“Guerriere;” but in this case the British court-martial subsequently
censured Captain Carden for mistakes. The battle lasted longer than
that with the “Guerriere,” and Decatur apologized for the extra hour
because the sea was high and his enemy had the weather-gauge
and kept at a distance; but the apology was not needed. Decatur
proved his skill by sparing his ship and crew. His own loss was
eleven men killed and wounded; the “Macedonian’s” loss was nine
times as great. The “United States” suffered little in her hull, and her
spars and rigging suffered no greater injury than could be quickly
repaired; while the “Macedonian” received a hundred shot in her
hull, and aloft nothing remained standing but her fore and main
masts and her fore-yard.
Decatur saved the “Macedonian,” and brought her back to New
London,—the only British frigate ever brought as a prize into an
American port. The two ships arrived December 4, and from New
London the “Macedonian” was taken to New York and received in
formal triumph. Captain Jones of the “Wasp” took command of her
in reward for his capture of the “Frolic.”
 Before the year closed, the “Constitution” had time for another
cruise. Hull at his own request received command of the Navy Yard
at Charlestown, and also took charge of the naval defences in New
York harbor, but did not again serve at sea during the war. The
“Constitution” was given to Captain Bainbridge, one of the oldest
officers in the service. A native of New Jersey, Bainbridge
commanded the “Philadelphia” when lost in the Tripolitan war, and
was held for eighteen months a prisoner in Tripoli. In 1812, when he
took command of the “Constitution,” though a year older than Hull
and five years older than Decatur, he had not yet reached his fortieth
year, while Rodgers, born in 1771, had but lately passed it. The
difference in age between these four naval officers and the four chief
generals—Dearborn, Wilkinson, Wade Hampton, and William Hull—
was surprising; for the average age of the naval commanders
amounted barely to thirty-seven years, while that of the four
generals reached fifty-eight. This difference alone accounted for
much of the difference in their fortune, and perhaps political
influence accounted for the rest.
Bainbridge showed no inferiority to the other officers of the
service, and no one grumbled at the retirement of Hull. The
“Constitution” sailed from Boston, October 25, with the “Hornet.”
The “Essex,” then in the Delaware, was ordered to join the squadron
at certain specified ports in the south Atlantic, and sailed October
28, expecting a very long cruise. December 13 Bainbridge arrived at
San Salvador, on the coast of Brazil, where he left the “Hornet” to
blockade the “Bonne Citoyenne,” a British 18-gun sloop-of-war
bound to England with specie. Cruising southward, within sight of
the Brazilian coast, in latitude 13° 6´ south, Bainbridge sighted the
British frigate “Java,” a ship of the same tonnage as the “Guerriere,”
throwing a slightly heavier broadside and carrying a large crew of
four hundred and twenty-six men, if the American account was
correct. Bainbridge tacked and made sail off shore, to draw the
stranger away from a neutral coast; the British frigate followed him,
until at half-past one o’clock in the afternoon Bainbridge shortened
sail, tacked again, and stood for his enemy. Soon after two o’clock
the action began, the two ships being on the same tack, the “Java”
to windward and the better sailer, and both fighting their long-range
guns. The British frigate insisted upon keeping at a distance, obliging
Bainbridge after half an hour to risk the danger of being raked; and
at twenty minutes before three o’clock the “Constitution” closed
within pistol-shot.[337] At ten minutes before three the ships were
foul, the “Java’s” jibboom in the “Constitution’s” mizzen rigging; and
from that point the battle became slaughter. In fifteen minutes the
“Java’s” bowsprit, fore-mast, and main top-mast were cut away, and
a few minutes after four o’clock she ceased firing. Her captain,
Lambert, was mortally wounded; the first lieutenant was wounded;
forty-eight of her officers and crew were dead or dying; one hundred
and two were wounded; little more than a hulk filled with wreck and
with dead or wounded men floated on the water.
The “Constitution” had but twelve men killed and twenty-two
wounded, and repaired damages in an hour. Owing perhaps to the
death of Captain Lambert the reports of the battle were more
contradictory than usual, but no one disputed that although the
“Java” was to windward and outsailed the American frigate, and
although her broadside counted as nearly nine against her enemy’s
ten,—for the “Constitution” on this cruise carried two guns less than
in her fight with the “Guerriere,”—yet the “Java” inflicted no more
damage than she ought to have done had she been only one fourth
the size of the American frigate, although she was defended more
desperately than either the “Guerriere” or the “Macedonian.”
With this battle the year ended. Bainbridge was obliged to blow
up his prize, and after landing and paroling his prisoners at San
Salvador sailed for Boston, where he arrived in safety, February 27,
1813. During the six months the war had lasted the little United
States navy captured three British frigates, besides the 20-gun
“Alert” and the 18-gun “Frolic;” privateers by scores had ravaged
British commerce, while the immense British force on the ocean had
succeeded only in capturing the little “Nautilus,” the 12-gun brig
“Vixen,” and the “Wasp.” The commerce of America had indeed
suffered almost total destruction; but the dispute was to be decided
not so much by the loss which England could inflict upon America, as
by that which America could inflict upon England.
 CHAPTER XVIII.
In such a war the people of the United States had only
themselves to fear; but their dangers were all the more formidable.
Had the war deeply disturbed the conditions of society, or brought
general and immediate distress, government and Union might easily
have fallen to pieces; but in the midst of military disaster and in
plain sight of the Government’s incompetence, the general public
neither felt nor had reason to fear much change in the routine of
life. Commerce had long accustomed itself to embargoes,
confiscations, and blockades, and ample supplies of foreign goods
continued to arrive. The people made no serious exertions; among a
population exceeding seven millions, not ten thousand men entered
the military service. The militia, liable to calls to the limit of one
hundred thousand, served for the most part only a few weeks in the
autumn, went home in whole regiments when they pleased,[338] and
in the East refused to go out at all. The scarcity of men was so great
that even among the sea-goingclass, for whose rights the war was
waged, only with the utmost difficulty and long delays, in spite of
bounties and glory, could sailors be found to man half-a-dozen
frigates for a three-months cruise, although the number of privateers
was never great.
The nation as a whole saw nothing of actual warfare. While
scarcely a city in Europe had escaped capture, and hardly a province
of that continent was so remote as not to be familiar with invading
armies or to have suffered in proportion to its resources, no
American city saw or greatly feared an enemy. The rich farms of
New York, New Jersey, Pennsylvania, and Virginia produced their
usual harvests, and except on exposed parts of the coast the
farmers never feared that their crops might be wasted by
manœuvring armies, or their cattle, pigs, and poultry be disturbed
by marauders. The country was vast, and quiet reigned throughout
the whole United States. Except at the little point of Niagara,
occupied by a few hundred scattered farmers, and on the extreme
outskirts of Ohio and Indiana, the occupations and industries of life
followed in the main their daily course.
The country refused to take the war seriously. A rich nation with
seven million inhabitants should have easily put one hundred
thousand men into the field, and should have found no difficulty in
supporting them; but no inducement that the Government dared
offer prevailed upon the people to risk life and property on a
sufficient scale in 1812. The ranks of the army were to be filled in
one of two ways,—either by enlistment in the regular service for five
years, with pay at five dollars a month, sixteen dollars bounty, and
on discharge three months pay and one hundred and sixty acres of
land; or by volunteer organizations to the limit of fifty thousand men
in all, officered under State laws, to serve for one year, with the pay
of regular troops but without bounty, clothed, and in case of cavalry
corps mounted, at their own expense. In a society where the day-
laborers’ wages were nowhere less than nine dollars a month,[339]
these inducements were not enough to supply the place of
enthusiasm. The patriotic citizen who wished to serve his country
without too much sacrifice, chose a third course,—he volunteered
under the Act of Congress which authorized the President to call one
hundred thousand State militia into service for six months; and upon
this State militia Dearborn, Hull, Van Rensselaer, and Smyth were
obliged chiefly to depend.
If the war fever burned hotly in any part of the country Kentucky
was the spot. There the whole male population was eager to prove
its earnestness. When Henry Clay returned to Lexington after the
declaration of war, he wrote to Monroe[340] that he was almost
alarmed at the ardor his State displayed; about four hundred men
had been recruited for the regular army, and although no one had
volunteered for twelve months, the quota of six-months militia was
more than supplied by volunteers.
“Such is the structure of our society, however,” continued Clay,
“that I doubt whether many can be engaged for a longer term than
 six months. For that term any force whatever which our population
can afford may be obtained. Engaged in agricultural pursuits, you are
well aware that from about this time, when the crop is either secured
in the barn or laid by in the field until the commencement of the
spring, there is leisure for any kind of enterprise.”
Clay feared only that these six-months militia corps, which had
armed and equipped themselves for instant service, might not be
called out. His friends were destined not to be disappointed, for
early in August pressing letters arrived from Hull’s army at Detroit
begging reinforcements, and the governor of Kentucky at once
summoned two thousand volunteers to rendezvous, August 20, at
Newport, opposite Cincinnati. This reinforcement could not reach
Detroit before the middle of September, and the difficulties already
developed in Hull’s path showed that the war could not be finished
in a single campaign of six months; but the Kentuckians were not on
that account willing to lengthen their term of service even to one
year.
The danger revealed by Hull’s position threw a double obstacle in
the way of public energy, for where it did not check, it promised to
mislead enthusiasm, and in either case it shook, if it did not destroy,
confidence in the national government. The leaders of the war party
saw their fears taking shape. Henry Clay wrote without reserve to
Monroe,[341]—
“Should Hull’s army be cut off, the effect on the public mind would
be, especially in this quarter, in the highest degree injurious. ‘Why did
he proceed with so inconsiderable a force?’ was the general inquiry
made of me. I maintained that it was sufficient. Should he meet with
a disaster, the predictions of those who pronounced his army
incompetent to its object will be fulfilled; and the Secretary of War, in
whom already there unfortunately exists no sort of confidence, cannot
possibly shield Mr. Madison from the odium which will attend such an
event.”
Clay was right in thinking that Eustis could not shield Madison;
but from the moment that Eustis could no longer serve that purpose,
Clay had no choice but to shield the President himself. When the
threatened disaster took place, victims like Eustis, Hull, Van
Rensselaer, Smyth, were sacrificed; but the sacrifice merely prepared
new material for other and perhaps worse disasters of the same
kind. In Kentucky this result was most strongly marked, for in their
irritation at the weakness of the national Government the
Kentuckians took the war into their own hands, appointed William
Henry Harrison to the command of their armies, and attempted to
conquer Canada by a campaign that should not be directed from
Washington. August 25 Clay described the feelings of his State by a
comparison suggesting the greatest military misfortunes known in
history:[342]
“If you will carry your recollections back to the age of the
Crusaders and of some of the most distinguished leaders of those
expeditions, you will have a picture of the enthusiasm existing in this
country for the expedition to Canada and for Harrison as commander.”

A week later, September 21, Clay gave another account, even

less assuring, of the manner in which the popular energy was
exhausting itself:—
“The capitulation of Detroit has produced no despair; it has, on the
contrary, awakened new energies and aroused the whole people of
this State. Kentucky has at this moment from eight to ten thousand
men in the field; it is not practicable to ascertain the precise number.
Except our quota of the hundred thousand militia the residue is chiefly
of a miscellaneous character, who have turned out without pay or
supplies of any kind, carrying with them their own arms and their own
subsistence. Parties are daily passing to the theatre of action; last
night seventy lay on my farm; and they go on, from a solitary
individual, to companies of ten, fifty, one hundred, etc. The only fear I
have is that the savages will, as their custom is, elude them, and
upon their return fall upon our frontiers. They have already shocked
us with some of the most horrid murders. Within twenty-four miles of
Louisville, on the headwaters of Silver Creek, twenty-two were
massacred a few days ago.”

The adventures of these volunteers made part of the next

campaign. Enthusiastic as Kentucky was, few or none of the eight or
ten thousand men under arms offered to serve for twelve months.
Excessively expensive, wasteful, insubordinate, and unsteady, no
general dared to depend on them. No one could be more conscious
of the evils of the system than the Government; but the Government
was helpless to invent a remedy.
“Proofs multiply daily,” wrote Madison to Monroe, September 21,
[343] “of the difficulty of obtaining regulars, and of the fluctuating
resource in the militia. High bounties and short enlistments, however
objectionable, will alone fill the ranks, and then too in a moderate
number.”

To dislike of prolonged service even the most ardent Western

supporters of the war added distrust of the Executive. The war
Republicans of the West and South were hardly less vigorous than
the Federalists of Massachusetts and Connecticut in their criticisms
of the Government at Washington. John Graham, chief clerk of the
State Department, who went to Kentucky in September, wrote to
Monroe[344] that “great as is the popularity of the President, it is
barely able to resist the torrent of public opinion against the
Secretary of War, who, so far as I can judge, is universally
considered by the people of this country as incompetent to his
present situation.” Clay’s opinion has already been shown; but the
angriest of all the war leaders on hearing of Hull’s surrender was
Senator Crawford of Georgia.
“Such is my want of confidence in the leaders of our forces,” he
wrote to Monroe,[345] “and their directors, Eustis and Hamilton in the
Cabinet, that I am fearful a continuance of the war, unless it should
be for several years, will only add to the number of our defeats. The
only difficulty I had in declaring war arose from the incompetency of
the men to whom the principal management of it was to be confided.
A Secretary of War who, instead of forming general and
comprehensive arrangements for the organization of his troops and
for the successful prosecution of the campaign, consumes his time in
reading advertisements of petty retailing merchants to find where he
may purchase one hundred shoes or two hundred hats; and a
Secretary of the Navy who, in instructing his naval officers, should
make the supply of the heads of departments with pineapples and
other tropical fruits through the exertions of these officers,—cannot
fail to bring disgrace upon themselves, their immediate employers,
and the nation. If Mr. Madison finds it impossible to bring his feelings
to consent to the dismission of unfaithful or incompetent officers, he
must be content with defeat and disgrace in all his efforts during the
 war. So far as he may suffer from this course he deserves no
commiseration, but his accountability to the nation will be great
indeed!”
Harsh as these comments, were, the Secretary of State found no
difficulty in listening to them; indeed, no member of the party was
more severe than Monroe. He visited Jefferson, and apparently
Jefferson agreed with his criticisms:[346]—
“We conferred on the then state of the Departments of War and
Navy, and agreed that whatever might be the merit of the gentlemen
in them, which was admitted in certain respects, a change in both was
indispensable.”
Indeed, Monroe did what no northern Democrat liked to do,—he
found fault with Dearborn.
“Our military operations,” he told Jefferson, “had been
unsuccessful. One army had been surrendered under circumstances
which impeached the integrity of the commander; and to the north, in
the whole extent of the country, so important and delicately
circumstanced as it was, the management had been most wretched.
The command at the important post of Niagara had been suffered to
fall into State hands, and to be perverted to local and selfish
purposes. Van Rensselaer, a weak and incompetent man with high
pretensions, took it. It was late in the year before General Dearborn
left Boston and repaired to Albany, and had given no impulse to the
recruiting business in the Eastern States by passing through them and
making appeals to the patriotism of the people; and when he took the
command at Albany it was in a manner to discourage all hope of
active operations during the favorable season. The commander ought
to lead every important movement. If intended to attack Montreal,
that being the grand attack, his station was there. If a smaller blow
only could be given, the feint against Montreal should have been
committed to another, while he commanded in person where real
service was to be performed. It was soon seen that nothing would be
done against Lower Canada; General Dearborn doubtless saw it on his
first arrival at Albany, if he did not anticipate it before he left Boston.
Niagara was the object next in importance, and had he taken the
command there he might and probably would, by superseding little
people and conducting our military operations, have prevented the
riotous and contentious scenes exhibited there, saved the country and
the Government from the disgraceful defeat of Van Rensselaer, and
 the more disgraceful and gasconading discomfiture of Smyth. The
experience of the campaign had excited a doubt with many, if not
with all, whether our military operations would prosper under General
Dearborn; ... he was advanced in years, infirm, and had given no
proof of activity or military talent during the year.”
The Secretary of State required nothing less than the retirement
of the two of his colleagues in the Cabinet, and of the general in
chief command of the army. The Secretary of the Treasury, though
less censorious than Clay, Crawford, or Monroe, shared their
opinions. He spoke of Eustis’s incompetence as a matter universally
admitted, and wrote to Jefferson that though the three disasters of
Hull, Van Rensselaer, and Smyth could not with justice be ascribed to
the Secretary of War, “yet his incapacity and the total want of
confidence in him were felt through every ramification of the public
service.”[347] Jefferson abstained from criticising the chief
incompetents, but set no bounds to his vindictiveness against the
unfortunate generals. “Hull will of course be shot for cowardice and
treachery,” he wrote to Madison;[348] “and will not Van Rensselaer
be broke for incapacity?”
The incapacity of Eustis, Hamilton, Dearborn, Hull, Van
Rensselaer, and Smyth pointed directly to the responsible source of
appointment,—the President himself; but in face of a general
election Republicans could not afford to criticise their President, and
only in private could they assail his Cabinet. The Federalists,
factious, weak, and unpopular as they were, expressed the secret
opinion of the whole country, and could be answered by no facts or
arguments except military success, which Madison was admittedly
incompetent to win; but perhaps the failure of his Cabinet, of his
generals, and of his troops gave the Federalists less advantage than
they drew from the failures of diplomacy in which his genius lay.
With reasons such as few nations ever waited to collect for an
appeal to arms, Madison had been so unfortunate in making the
issue that on his own showing no sufficient cause of war seemed to
exist. His management was so extraordinary that at the moment
when Hull surrendered Detroit, Great Britain was able to pose before
the world in the attitude of victim to a conspiracy between Napoleon
and the United States to destroy the liberties of Europe. Such
inversion of the truth passed ordinary bounds, and so real was
Madison’s diplomatic mismanagement that it paralyzed one half the
energies of the American people.
Largely if not chiefly owing to these mistakes, the New England
Federalists were able to convince themselves that Jefferson and
Madison were sold to France. From the moment war was declared,
the charge became a source of serious danger. Only one more step
was needed to throw the clerical party of New England into open
revolution. If the majority meant to close their long career by a
catastrophe which should leave the Union a wreck, they had but to
try the effects of coercion.
For a time the followers and friends of the Essex Junto had some
reason to hope that matters would quickly come to this pass, for the
declaration of war caused on both sides an outbreak of temper. In
Massachusetts, Governor Strong issued, June 26, a
proclamation[349] for a public Fast in consequence of the war just
declared “against the nation from which we are descended, and
which for many generations has been the bulwark of the religion we
profess;” and although such a description of England would in
previous times have scandalized the clergy, it was received with
general assent. The returning members of Congress who had voted
for war met a reception in some cases offensive and insulting, to the
point of actual assault. Two of the Massachusetts members, Seaver
and Widgery, were publicly insulted and hissed on Change in Boston;
while another, Charles Turner, member for the Plymouth district, and
Chief-Justice of the Court of Sessions for that county, was seized by
a crowd on the evening of August 3, on the main street of Plymouth,
and kicked through the town.[350] By energetic use of a social
machinery still almost irresistible, the Federalists and the clergy
checked or prevented every effort to assist the war, either by money
or enlistments. The Supreme Court of Massachusetts, with Chief-
Justice Parsons at its head, advised[351] Governor Strong that not to
Congress or to the President, but to the governor, belonged the right
to decide when the Constitutional exigency existed which should call
the State militia into the service of the United States; and Governor
Strong decided that neither foreign invasion nor domestic
insurrection existed, and that therefore he could not satisfy the
President’s request for the quota of the United States militia to
defend the coast. When, later in the season, the governor called out
three companies for the defence of Eastport and Castine, in Maine,
the chief-justice privately remonstrated, holding that this act yielded
the main point at issue between the State and National government.
[352]

General Dearborn’s annoyance at the difficulties thrown in the

way of enlistments was well-founded. By one favorite device, the
creation of fictitious debts, the person enlisting caused himself to be
arrested and bailed. The courts held that while the suit was pending
the man was the property of his bail, and could not be obliged to
resume his military duties.[353] Many such difficulties were created
by the activity of individuals; but organized efforts were made with
still more effect in counteracting the wishes of government. The
Federalist members of Congress issued an Address to their
constituents protesting against the action of Congress in suppressing
discussion; and this address declared the war to be unnecessary and
inexpedient. Immediately after the declaration, the House of
Representatives of Massachusetts issued another Address to the
People of the State,[354] declaring the war to be a wanton sacrifice
of their best interests, and asking their exertions to thwart it.
“To secure a full effect to your object, it will be necessary that you
should meet and consult together for the common good in your towns
and counties. It is in dark and trying times that this Constitutional
privilege becomes invaluable. Express your sentiments without fear,
and let the sound of your disapprobation of this war be loud and
deep. Let it be distinctly understood that in support of it your
conformity to the requisitions of law will be the result of principle and
not of choice. If your sons must be torn from you by conscription,
consign them to the care of God; but let there be no volunteers
except for defensive war.”
 The people at once acted upon the recommendation to hold
town-meetings and county conventions. Among the earliest was a
meeting in Essex County, July 21, Timothy Pickering presiding, which
adopted a declaration drawn by him, closing with his favorite
proposal of a State Convention, to which the meeting chose
delegates. This step—a revival of the old disunion project of 1804—
was received with general favor, and defeated only by the
courageous opposition of Samuel Dexter, who, breaking away from
his party associates, attacked the scheme so vigorously in Boston
town-meeting, August 6 and 7, that though Harrison Gray Otis and
other Federalists leaders gave it their public support, and though the
motion itself was carried, the plan was abandoned.[355]
Thenceforward, while towns and counties continued to adopt
addresses, memorials, and resolutions, they avoided committing
themselves to expressions or acts for which the time was not ripe. A
typical memorial among many that were showered upon the
President was adopted by a convention of electors of the county of
Rockingham in New Hampshire, August 5, and was the better worth
attention because drawn by Daniel Webster, who made there his first
appearance as a party leader:—
“We shrink from the separation of the States as an event fraught
with incalculable evils; and it is among our strongest objections to the
present course of measures that they have in our opinion a very
dangerous and alarming bearing on such an event. If a separation of
the States ever should take place, it will be on some occasion when
one portion of the country undertakes to control, to regulate, and to
sacrifice the interest of another; when a small and heated majority in
the government, taking counsel of their passions and not of their
reason, contemptuously disregarding the interests and perhaps
stopping the mouths of a large and respectable minority, shall by
hasty, rash, and ruinous measures threaten to destroy essential rights
and lay waste the most important interests. It shall be our most
fervent supplication to Heaven to avert both the event and the
occasion; and the Government may be assured that the tie that binds
us to the Union will never be broken by us.”

The conduct of England strengthened the Federalists. After the

repeal of the Orders in Council became known, Monroe, July 27,
authorized Jonathan Russell in London to arrange an armistice,
provided the British government would consent to an informal
arrangement in regard to impressments and blockades. Hardly had
these instructions been sent to England, when from Albany came
news that Sir George Prevost had proposed an armistice and General
Dearborn had accepted it. This act compelled the President either to
stop the war and disorganize his party, or to disapprove Dearborn’s
armistice without prejudice to the armistice which Russell was to
negotiate in London, and also without censure to General Dearborn.
To Dearborn the President, as the story has shown, sent immediate
orders for the renewal of hostilities; while Monroe, in fresh
instructions to Jonathan Russell,[356] explained the disavowal. The
explanations given by Monroe were little likely to satisfy Federalists
that the Government honestly wished for peace. Monroe alleged that
the repeal of the Orders in Council did not satisfy the United States,
because the repeal still asserted the principle underlying the orders,
which the United States could not admit; but he further maintained
that any armistice, made before obtaining redress on the subject of
impressments, might be taken as a relinquishment of the claim to
redress, and was therefore inadmissible.
However sound in principle these objections were, they seemed
to declare perpetual war; for until England should be reduced to the
position of Denmark or Prussia, she would not abandon in express
terms either the right of impressment or that of blockade. The
probable effect of a successful war waged on these grounds would
give Canada and the Floridas to the United States as the
consequence of aiding Napoleon to destroy European and English
liberties. The Federalist clergy had little difficulty in convincing their
congregations by such evidence that Madison was bound under
secret engagements with Napoleon; and while Madison planted
himself in the Napoleonic position of forcing war on a yielding
people, the British officials in Canada stood on the defensive,
avoided irritation, and encouraged trade and commerce. American
merchant-vessels carried British passes; and most of them, to the
anger of Napoleon, were freighted with supplies for the British army
in Portugal and Spain. The attitude of England would have been
magnificent in its repose had its dignity not been ruffled by the
conduct of Hull, Decatur, and Bainbridge, and by the privateers.
While the New England Federalists, taking the attitude of patriots
who strove only to avert impending ruin, made their profit of every
new national disaster, and repressed as well as they could the
indiscretions of their friends, the war party was not so well
disciplined. Democracies in history always suffered from the
necessity of uniting with much of the purest and best in human
nature a mass of ignorance and brutality lying at the bottom of all
societies. Although America was safe for the time from Old World
ruin, no political or military error went so far to disgust respectable
people with the war and its support, as an uprising of brutality which
occurred in Baltimore. Within some twenty years this newest of
American cities had gathered nearly fifty thousand inhabitants,
among whom were many of the roughest characters in America, fit
only for privateersmen or pirates, and familiar with both careers. On
the other hand, the State of Maryland like the State of Delaware
contained many conservatives, who showed their strength every four
years by depriving the Republican candidate for the Presidency of
some portion of the State’s electoral vote. Under their patronage a
newspaper called “The Federal Republican” was published in
Baltimore, edited by Jacob Wagner, who had been chief clerk of the
State Department under Secretary Pickering, and was retained in
that office by Secretary Madison until 1807, when he resigned the
place and made use of his knowledge to attack Madison in the press.
As an editor, Jacob Wagner belonged to the extreme wing of his
party, and scrupled at nothing in the way of an assertion or a
slander. His opposition to the war was bitter and unceasing, while
the city of Baltimore shared in the feeling common in the South and
West, that, after the declaration, opposition to the war amounted to
treason and should not be tolerated. June 22, immediately after the
declaration, a well-organized mob deliberately took possession of
Wagner’s printing office and destroyed it, pulling down even the
walls, while the citizens looked on and the mayor confined his
exercise of authority to deprecations.
Wagner removed to the District of Columbia, and began to
publish his paper in Georgetown, where the Government could be
made directly responsible in case of further violence; but his
associate, A. C. Hanson, and several of the Baltimore Federalists,
were not disposed to tolerate the dictation of a mob; and after
discussing the matter a month, some of them determined on an
attempt as fool-hardy as it was courageous.[357] Monday, July 27,
the “Federal Republican” was circulated among its subscribers in
Baltimore, purporting to be printed at 45 Charles Street, though
really printed at Georgetown; while about twenty persons, under the
general direction of Henry Lee,—a Virginian distinguished in the
Revolutionary War, and in 1791 governor of his State,—fortified
themselves in the house and waited attack. The same evening a
mob gathered and broke open the door. The garrison fired, and
killed or wounded some of the assailants. The attacking party
brought up a cannon, and a serious battle was about to begin, when
the mayor with a small squadron of cavalry intervened, and
persuaded Hanson and his friends to submit to the civil authority
and go to jail to answer for the blood they had shed. General Lee,
General Lingan,—also a Revolutionary officer,—Hanson, and the
other occupants of the house were marched to the jail through an
angry and violent mob. The city was in commotion, the authorities
were helpless, the militia when called upon did not appear; and that
night the mob, consisting chiefly of low Irish and Germans, entered
the jail and took out the prisoners. Some managed to escape in the
confusion; the rest were savagely beaten. Eight more or less
unconscious victims lay all night and till noon the next day piled on
the prison steps, and the crowd, which would not permit their
removal, amused itself by cutting and burning the sufferers to
ascertain whether they were dead. When at last the rioters
permitted them to be removed, General Lingan was in fact dead,
General Lee was crippled, and the others were more or less severely
injured.
 At that moment, and even long after the heat of temper
subsided,[358] party feeling tended to favor the rioters rather than
the Federalists, who had, as was said, “given aid and comfort to the
enemy;” but when the political effects of the massacre showed
themselves, the war party became aware that a blunder had been
committed more serious than any ordinary crime. The Baltimore
massacre recalled the excesses of the French Revolution, still fresh
in men’s minds; and although Democrats in Pennsylvania and
Republicans in Virginia might feel themselves too strong for disorder,
in the North and East the murder of Lingan shook the foundation of
society. Massachusetts and Connecticut looked to their arms. If their
political opinions were to be repressed by such means, they had
need to be unanimous on their own side. The town of Boston,
August 6, declared in strongly worded resolutions[359] that the riot
was “the first fruit of the unnatural and dreadful alliance into which
we have entered in fact, if not in form,” and ordered the magistrates
and citizens to be ready at a moment’s warning, armed and
equipped, to suppress any kind of disorder. Under this excitement,
the Federalists at Rockingham, August 5, talked of disunion, and the
rabble of Plymouth mobbed Turner on the night of August 3. If the
majority alone was to utter opinions, the Republican party north of
Pennsylvania might yet be forced to practise the virtue of silence.
Not all the political and military disasters of the year harmed the
Government and the war more seriously than they were injured by
the Baltimore mob.
Under the influence of such passions the Presidential election
approached. Except beyond the mountains the war party was
everywhere a social minority, and perhaps such strength as Madison
retained in the East consisted partly in the popular impression that
he was not a favorite with the authors of the war. The true
sentiment of the people, if capable of expression, was one of fretful
discontent; and the sense of diffused popular restlessness alone
explained the obstinacy of De Witt Clinton in refusing to desist from
his candidacy, and still more the first prominent appearance of
Martin Van Buren as manager of the intrigue for defeating Madison.