Professional Documents
Culture Documents
Statistics For Social Understanding With Stata and Spss Full Chapter
Statistics For Social Understanding With Stata and Spss Full Chapter
The idea for Statistics for Social Understand- more than sixty combined years of teach-
ing: With Stata and SPSS began with our ing experience in this area. We drew on
desire to offer a different kind of book to our teaching experience and research on
our statistics students. We wanted a book the teaching and learning of statistics
that would introduce students to the way to write what we think will be a more
statistics are actually used in the social effective textbook for fostering student
sciences: as a tool for advancing under- learning.
standing of the social world. We wanted In addition, we are excited to share
thorough coverage of statistical topics, our experiences teaching statistics to
with a balanced approach to calculation social science students by authoring the
and the use of statistical software, and book’s ancillary materials, which include
we wanted the textbook to cover the use not only practice problems, test banks,
of software as a way to explore data and and data sets but also suggested class ex-
answer exciting questions. We also wanted ercises, P owerPoint slides, assignments,
a textbook that incorporated Stata, which lecture notes, and class exercises.
is widely used in graduate programs and Statistics for Social Understanding is dis-
is increasingly used in undergraduate tinguished by several features: (1) It is the
classes, as well as SPSS, which remains only major introductory statistics book to
widespread. We wanted a book designed integrate Stata and SPSS, giving instruc-
for introductory students in the social sci- tors a choice of which software package
ences, including those with little quantita- to use. (2) It teaches statistics the way
tive background, but one that did not talk they are used in the social sciences. This
down to students and that covered the includes beginning every chapter with
conceptual aspects of statistics in detail examples from real research and taking
even when the mathematical details were students through research questions as
minimized. We wanted a clearly written, we cover statistical techniques or software
engaging book, with plenty of practice applications. It also includes extensive
problems of every type and easily avail- discussion of relationships between vari-
able data sets for classroom use. ables, through the earlier placement of the
We are excited to introduce this book chapter on cross-tabulation, the addition
to students and instructors. We are three of a dedicated chapter on causality, and
experienced instructors of statistics, two comparative examples throughout every
sociologists and a political scientist, with chapter of the book. (3) It is informed by
viii
Preface ix
doing, and the book provides numer- World Values Survey, and School Survey
ous opportunities for problem-solving. on Crime and Safety. Applied questions
• The book avoids the “busy” layout draw from sociology, political science,
used by some textbooks, which can criminology, and related fields. Several
distract students’ attention from the data sets, including all of those used in
content, particularly those with learn- the software sections, are available to stu-
ing differences. Drawing on the prin- dents and instructors (in both Stata and
ciples of universal design, our book SPSS formats) through the textbook’s
utilizes a clean, streamlined layout website. By using and making available
that will allow all students to focus on major social science data sets, we engage
the content without unnecessary dis- students in a problem-focused effort to
tractions.4 Boxes are clearly labeled make sense of real and engaging data
as either “In Depth,” which provide and enable them to ask and answer their
more detailed discussion or coverage own questions. Robust ancillary mate-
of more complex topics, or “Applica- rials, such as sample class exercises and
tion,” which provide additional exam- assignments, make it easy for instructors
ples. We avoid sidebars; terms defined to structure students’ engagement with
in the glossary are bolded and defined these data. The SPSS and Stata sections at
in the text, not in a sidebar. the end of each chapter allow students to
follow along.
• In keeping with principles of universal
Throughout the book, we discuss
design, we use both text and images to
issues and questions that working social
explain material (with more figures
scientists routinely confront, such as how
and illustrations than in many books).
to use missing data, recode variables
(including conceptual and statistical con-
3) Incorporates real-world siderations), combine variables into new
research and a real-world measures, think about outliers or atypi-
approach to the use of cal cases, choose appropriate measures,
statistics weigh considerations of causation, and
interpret results.
Each chapter begins with an engaging The focus in every chapter on relation-
real-world social science question and ships between variables or comparisons
examples from research. Chapters inte- across groups also reflects our commit-
grate examples and applications through- ment to showing students the power of
out. Chapters raise real-world questions statistics to answer important real-world
that can be addressed using a given tech- questions.
nique, explain the technique, provide an
example using the same question, and 4) Uses accessible, non-
show how related questions can also be condescending approach
addressed using Stata or SPSS. We use and tone
data sets that are widely used in the social
sciences, including the General Social Sur- We have written a text that is student-friendly
vey, American National Election Study, but not condescending. We have found that,
Preface xi
in an effort to assuage students’ anxiety for topics that are most important to
about statistics, some texts strike a tone that understanding statistical social sciences.
communicates the expectation that students Our aim is to give students the tools
lack confidence in their abilities. We are they might use as working researchers
conscious of the possibility that addressing in a variety of professions (from jobs in
students with the assumption that they hate small organizations where they might be
or are intimidated by statistics could acti- reading and writing up external data or
vate stereotype threat—the well-established doing program evaluation, to research
fact that, when students feel that they are or data analysis jobs) and prepare them
expected to perform poorly, their anxiety for higher-level statistics classes if they
over disproving that stereotype makes their choose to take them.
performance worse than it otherwise would
be. In selecting examples, we have remained
alert to the risk of stereotype threat, choos-
ing examples that do not activate (or even
For Instructors
challenge) gender or racial stereotypes Organization of the Text
about academic performance.
The textbook begins with descriptive sta-
tistics in chapters 2 through 5. One key dif-
5) Balances calculation
ference from many introductory statistics
and concepts
texts is that we introduce cross-tabulations
This book is aimed at courses that teach early, after frequency distributions and
statistics from the perspective of social before central tendency and variability.
science. Thus, the book frames the point In our experience as instructors, we have
of learning statistics as the analysis of noticed that students often begin think-
important social science questions. While ing about relationships between variables
we include some formulas and hand cal- at the very beginning of the class, asking
culation, we do so in order to help stu- questions about how groups differ in their
dents understand where the numbers frequency distributions of some variable,
come from. We believe students need to for example. Cross-tabulations follow nat-
be able to reason statistically, not simply urally at this point in the class and allow
use software to produce results, but we students to engage in real-world data anal-
recognize that most working research- ysis and investigate questions of causality
ers rely on statistical software, and we relatively early in the course. Chapters
strike a balance among these skills. At the 6 and 7 lay the foundation for inferential
same time, we spend more time on con- statistics, covering probability, the nor-
ceptual understanding, including more mal distribution, and sampling distribu-
in-depth consideration of topics relating tions. We cover elementary probability
to causality, and we include topics often in the context of the normal distribution,
omitted from other texts such as the use with a focus on the logic of probability
of confidence intervals as a follow-up to and probabilistic reasoning in order to lay
a hypothesis test. A lighter focus on hand the groundwork for an understanding of
calculation opens up time in the semester inferential statistics. Chapters 8 through
xii Preface
12 cover the basics of inferential statistics, don’t understand and ask questions or
including confidence intervals, hypothesis attend especially to that material in class.
testing, z- and t-tests, analysis of v ariance, After class, look back at the “Chapter
and chi-square. C hapter 13, unusual Summary” and work the practice prob-
among introductory statistics texts, focuses lems to consolidate your understanding.
on the logic of causality and control vari- If you found a chapter especially difficult
ables. Most existing texts address this topic on your first pass through, try to reread
more briefly (or not at all), but, in our expe- it after you have covered the material in
rience, it is an important topic that we all class. This may seem time-consuming, but
supplement in lecture. Finally, chapter 14 you not only will improve your under-
covers correlation and regression. While standing (and your grade) but will save
that chapter is pitched to an introductory time when it comes to studying for mid-
level, we pay more attention to multiple term and final exams or completing class
regression than do many texts, because it projects. As another student explained:
is so widely used, and we have a box on
logistic regression to introduce students The textbook format let me go through
to the range of models that working social the material from class at a slower
scientists employ. pace and I could turn to it for step-by-
step help in doing the assignments.
Instructors who wish to cover chap-
ters in a different order—for example,
Similarly, you should look through the
delaying cross-tabulations until later in
software sections before you conduct
the semester—can readily do so. Some
these exercises in class or lab. You do not
courses may not cover probability or
need to try to memorize the SPSS or Stata
analysis of variance, and those chapters
commands, but familiarize yourself with
can be omitted. For instructors who want
the procedures and the reasons for them.
to follow the order of this book in their
As with the rest of the chapter, hands-on
class, the ancillary materials make it easy
practice is key here, too.
to do so.
Remember, you are taking this class
because you want to understand the social
world. As another of our students wrote:
For Students
If you are not too familiar with work-
In a course evaluation, one of our students ing with numbers, that is just fine!
offered advice to future students: This course is designed as an analyt-
Use the textbook! it is incredibly specific ical course which means that you will
and helpful. be focusing more so on the meaning
behind numbers and statistics rather
We agree, and not just because we wrote than just focusing on finding “correct”
it! We suggest reading the assigned sec- answers.
tion of the chapter before class and work-
ing the example problems, pencil in hand, The companion website contains more
as you read. Make a note of anything you study materials and gives you access to
Preface xiii
the data sets used for the software sections PowerPoint® Slides. The PowerPoint
in the textbook. You can use these data sets presentation provides lecture slides for
and your newfound skill in SPSS or Stata every chapter. In addition, multiple choice
to investigate questions you are interested review slides for classroom use are avail-
in, beyond those we cover. able for each chapter. The presentation is
Chapter 1 contains more tips on study- available to adopters for download on the
ing and learning as well as overcoming text’s catalog page at https://rowman.
math anxiety. com/ISBN/9781538109830.
For Students
Ancillaries
Companion Website. Accompanying the
This book is accompanied by a learning text is an open-access Companion Website
package, written by the authors, that is designed to reinforce key topics and con-
designed to enhance the experience of
cepts. For each chapter, students will have
both instructors and students. access to:
Publicly available data sets used in the
For Instructors Stata and SPSS sections
Instructor’s Manual with Solutions. Flashcards of key concepts
This valuable resource includes a sam-
Discussion questions
ple course syllabus and links to the pub-
licly available data sets used in the Stata Students can access the Companion
and SPSS sections of the text. For each Website from their computers or mobile
chapter, it includes lecture notes, sug- devices at https://textbooks.rowman.
gested classroom activities, discussion com/whittier.
questions, and the solutions to the prac-
tice problems. The Instructor’s Manual
with Solutions is available to adopt- Acknowledgements
ers for download on the text’s catalog
page at https://rowman.com/ISBN/ We are grateful to many manuscript
9781538109830. reviewers, both those who are identified
here and those who chose to remain anon-
Test Bank. The Test Bank includes both ymous, for their in-depth and thoughtful
short answer and multiple choice items comments as we developed this text. We
and is available in either Word or Respon- are fortunate to have benefited from their
dus format. In either format, the Test Bank knowledgeable and helpful input. We
can be fully edited and customized to best thank the following reviewers:
meet your needs. The Test Bank is avail-
able to adopters for download on the text’s Jacqueline Bergdahl, Department of
catalog page at https://rowman.com/ Sociology and Anthropology, Wright
ISBN/9781538109830. State University
xiv Preface
Notes
1
S. Friel. 2007. “The Research Frontier: Where Tech- 3
J. Garfield and D. Ben-Zvi. 2007. “How Students Learn
nology Interacts with the Teaching and Learning of Statistics Revisited: A Current Review of Research on
Data Analysis.” In M. K. Heid and G. W. Blume (eds.), Teaching and Learning Statistics.” International Statisti-
Research on Technology and the Teaching and Learn- cal Review 75: 372–396.
ing of Mathematics: Syntheses and Perspectives, 4
S. E. Burgstahler. 2015. Universal Design in Higher
Volume 2 (pp. 279–331). Greenwich: Information Age Education: From Principles to Practice. Cambridge, MA:
Publishing, Inc. Harvard Education Press.
2
J. R. Star. 2016. “Small Steps Forward: Improving Math- 5
S. J. Spencer, C. Logel, and P. G. Davies. 2016.
ematics Instruction Incrementally.” Phi Delta Kappan 97: “Stereotype Threat.” Annual Review of Psychology 67:
58–62. 415–437.
About the Authors
xvi
Introduction
Chapter 1 Using Statistics to Study the
Social World
1
2 CHAPTER 1 Introduction
Statistics and quantitative data are important tools for understanding large-scale
social and political processes and institutions as well as how these structures shape
individual lives. They help us to comprehend trends and patterns that are too large for
us to see in other ways. Statistics do this in three main ways. First, they help us simply
to describe large-scale patterns. For example, what is the average income of residents in
a given state? Second, statistics help us determine the factors that shape these patterns.
This includes simple comparisons, such as how income varies by gender or by age.
It also includes more complicated mathematical models that can show how multiple
forces shape a given outcome. How do gender, age, race, and education interact to
shape income, for example? Third, statistics help us understand how and whether we
can generalize from data gathered from only some members of a group to draw con-
clusions about all members of that group. This aspect of statistics, called inferential sta-
tistics, uses ideas about probability to determine what kinds of generalizations we can
make. It is what allows researchers to draw meaningful conclusions from data about
relatively small numbers of people.
In this book, we emphasize what we can do with statistics, focusing on real social
science research and analyzing real data. Readers of this book will develop a strong
sense of how quantitative social scientists conduct their research and will get plenty
of practice in analyzing social science data. Not all of this book’s readers will pursue
careers as researchers, but many of you will have careers that include analyzing and
presenting information. And, all of you face the task of making sense of mountains
of information, including social science research findings, communicated by various
media. This book provides essential tools for doing so.
Recently, some commentators have noted that we have entered a “post-fact,” or
“post-truth,” era. People mean different things by this, but one meaning is that the
sheer volume of people and agencies producing facts has multiplied to the point that
an expert can be found to attest to the accuracy of just about any claim.1 Just think of
the amount of information that you are exposed to on a weekly basis from various
social media platforms, websites, television, and other forms of media. How do you
make sense of it? How do you, for example, decide whether a claim you read online is
true or false? Statistics can powerfully influence opinion because they use numerical
data, which American culture assumes are objective and legitimate. But not all claims
are equally factual, even those that appear to be backed up by statistics. This book will
equip you with an understanding of how statistics work so that you can evaluate the
meaning and credibility of statistical data for yourself.
When quantitative research is carefully conceived and conducted, the results
of statistical analyses can yield valuable information not only about how the social
world works but also about how to effectively address social problems. For example,
in her 2007 book Marked, sociologist Devah Pager examined how having a criminal
record affects men’s employment prospects in blue collar jobs.2 She conducted a study
in which she hired paid research assistants, called testers, to submit fake résumés in
person to potential employers. The résumés were the same, with the only difference
Research Questions and the Research Process 3
being that some of them listed a parole officer as a reference, indicating that the appli-
cant had spent time in prison, while the others did not have a parole officer as a ref-
erence. Did résumés without the parole officer reference fare better in the job search
process? Yes, they did. On average former offenders were 46% less likely to receive
a callback about the job, and the results of the analysis suggested that this difference
could be generalized to the overall population of men applying for blue collar jobs,
not just the testers in her study.3 Pager also varied the race of the testers applying
for jobs—half were white, and half were black. She found that having the mark of a
criminal record reduced the chances of a callback by 64% for black testers and 50%
for white testers, indicating that the damage of a criminal record is particularly acute
for black men.
By varying only whether the applicant had a criminal record, Pager controlled for
alternative explanations of the negative effect of a criminal record on the likelihood of
receiving a callback for a job. In other words, employers were reacting to the criminal
record itself, not factors that might be associated with a criminal record, such as erratic
work histories.
Pager’s study contains many of the key elements of statistical analysis that we dis-
cuss in this book: assessment of the relationship between two variables (criminal record
and employer callbacks); a careful investigation of whether one of the variables (crimi-
nal record) has a causal impact on the other (employer callback) and, if so, whether that
causal impact varies by another factor (race); and examination of the generalizability of
the results.
that sponsor them, whether they include violence, as well as their size, in order to
explore what shapes media coverage. Statistical methods can support both approaches
to research.
This book focuses on quantitative analysis—that is, analyses that use statistical
techniques to analyze numerical data. Many social scientists also use qualitative meth-
ods. Qualitative methods start with data that are not numerical, such as the text of
documents, interviews, or field observations. Qualitative data analysis often focuses on
meanings, processes, and interactions; like quantitative research, it may test hypotheses
or be more exploratory in nature. Qualitative research analysis often uses specialized
software programs. Increasingly, many researchers use mixed methods, which employ
both qualitative and quantitative data and analysis. While this book focuses on quan-
titative analysis, combining both methods can yield a richer and more accurate under-
standing of social phenomena than either approach alone.
Concept
Theorecal
Well-being
Definion:
Sense of overall, health, sasfacon, and
comfort in life
Dimensions
Variables
Empirical
Frequency Number Rang of Frequency Sense of Sense of
Stress View of
of sick healthy of meaning purpose
of exercise level self
days eang habits depression in life in life
predefined response categories. The number of categories can range from as little as
two (e.g., yes or no) to very many (e.g., a feeling thermometer that asks respondents to
rate their feeling about something on a scale from 0 to 100 degrees). With closed-ended
survey items, the researcher decides on the measurement of the concept before admin-
istering the survey. Open-ended survey items do not provide response categories. For
example, an item might ask respondents to name the issue that is most important to
them in casting a vote for a candidate. Open-ended items give respondents more lee-
way in answering questions. Once the researcher has all responses to an open-ended
item, the researcher often devises response categories informed by the responses them-
selves and then assigns respondents to those categories based on their responses. For
example, with an open-ended question about which issues are important to voters, the
researcher might combine various responses having to do with jobs or the economy
into one category.
Units of Analysis
In the social sciences, researchers are interested in studying the characteristics of
individuals but also the characteristics of groups. Who or what is being studied is
the unit of analysis. A study of people’s voting patterns and political party affiliation
focuses on understanding individuals. But a study of counties that voted for a
Republican vs. Democratic candidate focuses on understanding characteristics of a
group, in this case counties. In the first case, researchers might seek to understand
what explains people’s votes; in the second case, researchers might seek to understand
what characteristics are associated with Republican vs. Democratic counties. When
the unit of measurement is the group, we sometimes also refer to it as aggregate level.
Aggregate-level units that researchers might be interested in include geographic
areas, organizations, religious congregations, families, sports teams, musical groups,
or businesses. One must be careful about making inferences across different levels
of measurement. A county may be Republican, but at the individual level, there are
both Democratic and Republican residents of that county. Drawing conclusions about
individuals based on the groups to which they belong is an error in logic known as the
ecological fallacy.
track of the income produced by each of them. Still others might purposefully report a
number that is higher or lower than their actual income. Researchers never know for
sure how much error their variables contain, but we can evaluate and minimize error in
measurement by assessing the validity and reliability of our variables.
Validity indicates the extent to which variables actually measure what they claim
to measure. When measures have a high degree of validity, this means that there is a
strong connection between the measurement of a concept and its conceptual definition.
In other words, valid measures are accurate indicators of the underlying concept. Imag-
ine a researcher who claims that he has found that happiness declines as people exercise
more. How is that researcher measuring happiness? It turns out that he has operational-
ized happiness through responses to two survey questions: “How much energy do you
feel you have?” and “How much do you look forward to participating in family activ-
ities?” Do you think answers to these questions are good measures of happiness? They
may get at elements of happiness—happier people may have more energy or look for-
ward to participating in activities more. But they are not direct measures of happiness,
and we could argue that they measure other things instead (such as how busy people
are or their health). What about a researcher who wants to measure the prevalence of
food insecurity, in which people do not have consistent access to sufficient food? This
could be operationalized in a survey question such as, “How often do you have insuf-
ficient food for yourself and your family” or “How often do you go hungry because
of inability to get sufficient food for yourself or your family?” It could also be oper-
ationalized by the number and size of food pantries per capita or food stamp usage.
Which way of operationalizing food insecurity is more accurate? The survey questions
have greater validity because both food pantries and food stamp usage are affected by
forces other than food insecurity (urban areas may have more food pantries per capita
than rural areas, not all people eligible for food stamps use them, and so forth). If the
researcher were interested instead in social services to reduce food insecurity, looking
at food pantries and food stamps would be a valid measure.
Even if a measure is valid, it may not yield consistent answers. This is the ques-
tion of reliability. Reliable measures are those whose values are unaffected by the
measurement process or the measurement instrument itself (e.g., the survey). Imagine
asking the same group of college students to rate how often in a typical week they
spend time with friends, with the following response choices: “often,” “a few times,”
“occasionally,” and “rarely.” These response choices are likely to lead to problems with
reliability, because they are not precise. A student who gets together with friends about
five times a week might choose “often” or “a few times,” and if you asked her the
question again a week later she might choose the other option, even if her underlying
estimate of how often she spent time with friends was unchanged. In other words, the
same students may give quite different, or inconsistent, responses if asked the ques-
tion repeatedly.
Measures also tend not to be reliable when they ask questions that respondents may
not have detailed understanding or information about. For example, a survey might ask
how many minutes a week people spend doing housework, or a survey of Americans
8 CHAPTER 1 Introduction
might ask their opinion of Britain’s foreign policy toward Chile. Because people do not
generally precisely track minutes spent doing housework, and Americans are unlikely
to know much about British foreign policy, their responses to such questions will be
inconsistent.
Reliability and validity do not necessarily coincide. For example, the time shown on
a clock may be reliable without being valid. Some households may deliberately set their
clocks to be a few minutes fast, ensuring that when the alarm goes off at what the clock
says is 6:45, the actual time is 6:30. In this case, the clock consistently—that is, reliably—
tells time, but that time is always wrong (or invalid).
Figure 1.2 uses a feeling thermometer, which asks people to rate their feeling about
something on a scale from 0 to 100 degrees, to illustrate how reliability and validity
can coincide or not. Imagine these are an individual’s responses to the same feeling
thermometer item asked five separate times. The true value of the person’s feeling is
42 degrees. In scenario A, the responses have a high degree of validity, or accuracy,
because they are all near 42 degrees, the accurate value. There is also a high degree of
reliability because the responses are consistent. Researchers strive to attain scenario A
by obtaining accurate and consistent measures. In scenario B, there is still a high degree
of consistency, and therefore reliability, in the measure. However, validity is low because
the responses are far from the true value of 42 degrees. Finally, scenario C reflects both
low reliability and low validity. The responses are inconsistent, or scattered across the
A. High Reliability, High Validity B. High Reliability, Low Validity C. Low Reliability, Low Validity
50 50 50
True value: 42 True value: 42 True value: 42
0 0 0
These words would be sung over and over again, with others,
improvised as they went along—jargon, perhaps, to the reader, but
full of meaning to the singers. I have sometimes thought that the
mere hearing of these songs would have done more to impress the
good people of the north with the soul-crushing character of slavery
than whole volumes exposing the physical cruelties of the slave
system; for the heart has no language like song. Many years ago,
when recollecting my experience in this respect, I wrote of these
slave songs in the following strain:
“I did not, when a slave, fully understand the deep meaning of
those rude and apparently incoherent songs. I was, myself, within
the circle, so that I could then neither hear nor see as those without
might see and hear. They breathed the prayer and complaint of souls
overflowing with the bitterest anguish. They depressed my spirits
and filled my heart with ineffable sadness.”
The remark in the olden time was not unfrequently made, that
slaves were the most contented and happy laborers in the world, and
their dancing and singing were referred to in proof of this alleged
fact; but it was a great mistake to suppose them happy because they
sometimes made those joyful noises. The songs of the slaves
represented their sorrows, rather than their joys. Like tears, they
were a relief to aching hearts. It is not inconsistent with the
constitution of the human mind, that avails itself of one and the same
method for expressing opposite emotions. Sorrow and desolation
have their songs, as well as joy and peace.
It was the boast of slaveholders that their slaves enjoyed more of
the physical comforts of life than the peasantry of any country in the
world. My experience contradicts this. The men and the women
slaves on Col. Lloyd’s farm received as their monthly allowance of
food, eight pounds of pickled pork, or its equivalent in fish. The pork
was often tainted, and the fish were of the poorest quality. With their
pork or fish, they had given them one bushel of Indian meal,
unbolted, of which quite fifteen per cent. was more fit for pigs than
for men. With this one pint of salt was given, and this was the entire
monthly allowance of a full-grown slave, working constantly in the
open field from morning till night every day in the month except
Sunday. There is no kind of work which really requires a better
supply of food to prevent physical exhaustion than the field work of a
slave. The yearly allowance of clothing was not more ample than the
supply of food. It consisted of two tow-linen shirts, one pair of
trowsers of the same coarse material, for summer, and a woolen pair
of trowsers and a woolen jacket for winter, with one pair of yarn
stockings and a pair of shoes of the coarsest description. Children
under ten years old had neither shoes, stockings, jackets, nor
trowsers. They had two coarse tow-linen shirts per year, and when
these were worn out they went naked till the next allowance day—
and this was the condition of the little girls as well as the boys. As to
beds, they had none. One coarse blanket was given them, and this
only to the men and women. The children stuck themselves in holes
and corners about the quarters, often in the corners of huge
chimneys, with their feet in the ashes to keep them warm. The want
of beds, however, was not considered a great privation by the field
hands. Time to sleep was of far greater importance. For when the
day’s work was done most of these had their washing, mending, and
cooking to do, and having few or no facilities for doing such things,
very many of their needed sleeping hours were consumed in
necessary preparations for the labors of the coming day. The
sleeping apartments, if they could have been properly called such,
had little regard to comfort or decency. Old and young, male and
female, married and single, dropped down upon the common clay
floor, each covering up with his or her blanket, their only protection
from cold or exposure. The night, however, was shortened at both
ends. The slaves worked often as long as they could see, and were
late in cooking and mending for the coming day, and at the first gray
streak of the morning they were summoned to the field by the
overseer’s horn. They were whipped for over-sleeping more than for
any other fault. Neither age nor sex found any favor. The overseer
stood at the quarter door, armed with stick and whip, ready to deal
heavy blows upon any who might be a little behind time. When the
horn was blown there was a rush for the door, for the hindermost one
was sure to get a blow from the overseer. Young mothers who
worked in the field were allowed an hour about ten o’clock in the
morning to go home to nurse their children. This was when they
were not required to take them to the field with them, and leave them
upon “turning row,” or in the corner of the fences.
As a general rule the slaves did not come to their quarters to
take their meals, but took their ash-cake (called thus because baked
in the ashes) and piece of pork, or their salt herrings, where they
were at work.
But let us now leave the rough usage of the field, where vulgar
coarseness and brutal cruelty flourished as rank as weeds in the
tropics, where a vile wretch, in the shape of a man, rides, walks, and
struts about, with whip in hand, dealing heavy blows and leaving
deep gashes on the flesh of men and women, and turn our attention
to the less repulsive slave life as it existed in the home of my
childhood. Some idea of the splendor of that place sixty years ago
has already been given. The contrast between the condition of the
slaves and that of their masters was marvelously sharp and striking.
There were pride, pomp, and luxury on the one hand, servility,
dejection, and misery on the other.
CHAPTER VII.
LUXURIES AT THE GREAT HOUSE.
THE close-fisted stinginess that fed the poor slave on coarse corn-
meal and tainted meat, that clothed him in crashy tow-linen and
hurried him on to toil through the field in all weathers, with wind and
rain beating through his tattered garments, that scarcely gave even
the young slave-mother time to nurse her infant in the fence-corner,
wholly vanished on approaching the sacred precincts of the “Great
House” itself. There the scriptural phrase descriptive of the wealthy
found exact illustration. The highly-favored inmates of this mansion
were literally arrayed in “purple and fine linen, and fared
sumptuously every day.” The table of this house groaned under the
blood-bought luxuries gathered with painstaking care at home and
abroad. Fields, forests, rivers, and seas were made tributary.
Immense wealth and its lavish expenditures filled the Great House
with all that could please the eye or tempt the taste. Fish, flesh, and
fowl were here in profusion. Chickens of all breeds; ducks of all
kinds, wild and tame, the common and the huge Muscovite; Guinea
fowls, turkeys, geese, and pea-fowls were fat, and fattening for the
destined vortex. Here the graceful swan, the mongrels, the black-
necked wild goose, partridges, quails, pheasants and pigeons,
choice water-fowl, with all their strange varieties, were caught in this
huge net. Beef, veal, mutton, and venison, of the most select kinds
and quality, rolled in bounteous profusion to this grand consumer.
The teeming riches of the Chesapeake Bay, its rock perch, drums,
crocus, trout, oysters, crabs, and terrapin were drawn hither to adorn
the glittering table. The dairy, too, the finest then on the eastern
shore of Maryland, supplied by cattle of the best English stock,
imported for the express purpose, poured its rich donations of
fragrant cheese, golden butter, and delicious cream to heighten the
attractions of the gorgeous, unending round of feasting. Nor were the
fruits of the earth overlooked. The fertile garden, many acres in size,
constituting a separate establishment distinct from the common farm,
with its scientific gardener direct from Scotland, a Mr. McDermott,
and four men under his direction, was not behind, either in the
abundance or in the delicacy of its contributions. The tender
asparagus, the crispy celery, and the delicate cauliflower, egg plants,
beets, lettuce, parsnips, peas, and French beans, early and late,
radishes, cantelopes, melons of all kinds; and the fruits of all climes
and of every description, from the hardy apples of the north to the
lemon and orange of the south, culminated at this point. Here were
gathered figs, raisins, almonds, and grapes from Spain, wines and
brandies from France, teas of various flavor from China, and rich,
aromatic coffee from Java, all conspiring to swell the tide of high life,
where pride and indolence lounged in magnificence and satiety.
Behind the tall-backed and elaborately wrought chairs stood the
servants, fifteen in number, carefully selected, not only with a view to
their capacity and adeptness, but with especial regard to their
personal appearance, their graceful agility, and pleasing address.
Some of these servants, armed with fans, wafted reviving breezes to
the over-heated brows of the alabaster ladies, whilst others watched
with eager eye and fawn-like step, anticipating and supplying wants
before they were sufficiently formed to be announced by word or
sign.
These servants constituted a sort of black aristocracy. They
resembled the field hands in nothing except their color, and in this
they held the advantage of a velvet-like glossiness, rich and
beautiful. The hair, too, showed the same advantage. The delicately-
formed colored maid rustled in the scarcely-worn silk of her young
mistress, while the servant men were equally well attired from the
overflowing wardrobe of their young masters, so that in dress, as
well as in form and feature, in manner and speech, in tastes and
habits, the distance between these favored few and the sorrow and
hunger-smitten multitudes of the quarter and the field was immense.
In the stables and carriage-houses were to be found the same
evidences of pride and luxurious extravagance. Here were three
splendid coaches, soft within and lustrous without. Here, too, were
gigs, phaetons, barouches, sulkeys, and sleighs. Here were saddles
and harnesses, beautifully wrought and richly mounted. Not less
than thirty-five horses of the best approved blood, both for speed
and beauty, were kept only for pleasure. The care of these horses
constituted the entire occupation of two men, one or the other of
them being always in the stable to answer any call which might be
made from the Great House. Over the way from the stable was a
house built expressly for the hounds, a pack of twenty-five or thirty,
the fare for which would have made glad the hearts of a dozen
slaves. Horses and hounds, however, were not the only consumers
of the slave’s toil. The hospitality practiced at the Lloyd’s would have
astonished and charmed many a health-seeking divine or merchant
from the north. Viewed from his table, and not from the field, Colonel
Lloyd was, indeed, a model of generous hospitality. His house was
literally a hotel for weeks, during the summer months. At these
times, especially, the air was freighted with the rich fumes of baking,
boiling, roasting, and broiling. It was something to me that I could
share these odors with the winds, even if the meats themselves were
under a more stringent monopoly. In master Daniel I had a friend at
court, who would sometimes give me a cake, and who kept me well
informed as to their guests and their entertainments. Viewed from
Col. Lloyd’s table, who could have said that his slaves were not well
clad and well cared for? Who would have said they did not glory in
being the slaves of such a master? Who but a fanatic could have
seen any cause for sympathy for either master or slave? Alas, this
immense wealth, this gilded splendor, this profusion of luxury, this
exemption from toil, this life of ease, this sea of plenty were not the
pearly gates they seemed to a world of happiness and sweet
content. The poor slave, on his hard pine plank, scantily covered
with his thin blanket, slept more soundly than the feverish voluptuary
who reclined upon his downy pillow. Food to the indolent is poison,
not sustenance. Lurking beneath the rich and tempting viands were
invisible spirits of evil, which filled the self-deluded gourmandizer
with aches and pains, passions uncontrollable, fierce tempers,
dyspepsia, rheumatism, lumbago, and gout, and of these the Lloyds
had a full share.