Вы находитесь на странице: 1из 15

Sociological research is the scientific means of acquiring information about various

aspects of society and social behaviour.

A goal of sociological research is to discover the similarities, differences, patterns, and trends
of a given population.
The best and most representative sample is a random sample, in which each member of a
population has an equal chance of being chosen as a subject.
In quantitative research, information collected from respondents (for example, a
respondent's college ranking) is converted into numbers (for example, a junior may equal
three and a senior four).
In qualitative research, information collected from respondents takes the form of verbal
descriptions or direct observations of events. Although verbal descriptions and observations
are useful, many scientists prefer quantitative data for purposes of analysis.
To analyse data, scientists use statistics, which is a collection of mathematical procedures for
describing and drawing inferences from the data. Two types of statistics are most
common: inferential, used for making predictions about the population, and descriptive,
used for describing the characteristics of the population and respondents. Scientists use both
types of statistics to draw general conclusions about the population being studied and the
A scientist who uses a questionnaire or test in a study is interested in the test's validity, which
is its capacity to measure what it purports to measure. He or she is also interested in
its reliability, or capacity to provide consistent results when administered on different

Sociological Research: Designs, Methods

Sociologists use many different designs and methods to study society and social behaviour.
Most sociological research involves ethnography, or field work designed to depict the
characteristics of a population as fully as possible.

Three popular social research designs (models) are

Crosssectional, in which scientists study a number of individuals of different ages
who have the same trait or characteristic of interest at a single time

Longitudinal, in which scientists study the same individuals or society repeatedly over
a specified period of time

Crosssequential, in which scientists test individuals in a crosssectional sample more

than once over a specified period of time

Six of the most popular sociological research methods (procedures) are the case study,
survey, observational, correlational, experimental, and crosscultural methods, as well as
working with information already available.

Case study research

In case study research, an investigator studies an individual or small group of individuals
with an unusual condition or situation. Case studies are typically clinical in scope. The
investigator (often a clinical sociologist) sometimes uses selfreport measures to acquire
quantifiable data on the subject. A comprehensive case study, including a longterm follow
up, can last months or years.
On the positive side, case studies obtain useful information about individuals and small
groups. On the negative side, they tend to apply only to individuals with similar
characteristics rather than to the general population. The high likelihood of the investigator's
biases affecting subjects' responses limits the generalizability of this method.

Survey research
Survey research involves interviewing or administering questionnaires, or written surveys,
to large numbers of people. The investigator analyses the data obtained from surveys to learn
about similarities, differences, and trends. He or she then makes predictions about the
population being studied.
As with most research methods, survey research brings both advantages and disadvantages.
Advantages include obtaining information from a large number of respondents, conducting
personal interviews at a time convenient for respondents, and acquiring data as inexpensively
as possible. Mailin surveys have the added advantage of ensuring anonymity and thus
prompting respondents to answer questions truthfully.
Disadvantages of survey research include volunteer bias, interviewer bias, and
distortion. Volunteer bias occurs when a sample of volunteers is not representative of the
general population. Subjects who are willing to talk about certain topics may answer surveys
differently than those who are not willing to talk. Interviewer bias occurs when an
interviewer's expectations or insignificant gestures (for example, frowning or smiling)
inadvertently influence a subject's responses one way or the other. Distortion occurs when a
subject does not respond to questions honestly.

Observational research
Because distortion can be a serious limitation of surveys, observational research involves
directly observing subjects' reactions, either in a laboratory (called laboratory observation)
or in a natural setting (called naturalistic observation). Observational research reduces the
possibility that subjects will not give totally honest accounts of the experiences, not take the
study seriously, and fail to remember, or feel embarrassed.

Observational research has limitations, however. Subject bias is common, because volunteer
subjects may not be representative of the general public. Individuals who agree to observation
and monitoring may function differently than those who do not. They may also function
differently in a laboratory setting than they do in other settings.

Correlational research
A sociologist may also conduct correlational research. A correlation is a relationship

between two variables (or factors that change). These factors can be characteristics,
attitudes, behaviours, or events. Correlational research attempts to determine if a relationship
exists between the two variables, and the degree of that relationship.
A social researcher can use case studies, surveys, interviews, and observational research to
discover correlations. Correlations are either positive (to +1.0), negative (to 1.0), or nonexistent (0.0). In a positive correlation, the values of the variables increase or decrease (co
vary) together. In a negative correlation, one variable increases as the other decreases. In a
non-existent correlation, no relationship exists between the variables.
People commonly confuse correlation with causation. Correlational data do not
indicate causeandeffect relationships. When a correlation exists, changes in the value of one
variable reflect changes in the value of the other. The correlation does not imply that one
variable causes the other, only that both variables somehow relate to one another. To study
the effects that variables have on each other, an investigator must conduct an experiment.

Experimental research
Experimental research attempts to determine how and why something happens. Experimental
research tests the way in which an independent variable (the factor that the scientist
manipulates) affects a dependent variable (the factor that the scientist observes).
A number of factors can affect the outcome of any type of experimental research. One is
finding samples that are random and representative of the population being studied. Another
is experimenter bias, in which the researcher's expectations about what should or should not
happen in the study sway the results. Still another is controlling for extraneous variables,
such as room temperature or noise level, that may interfere with the results of the experiment.
Only when the experimenter carefully controls for extraneous variables can she or he draw
valid conclusions about the effects of specific variables on other variables.

Cross-cultural research
Sensitivity to others' norms, folkways, values, mores, attitudes, customs, and practices
necessitates knowledge of other societies and cultures. Sociologists may conduct cross
cultural research, or research designed to reveal variations across different groups of people.
Most crosscultural research involves survey, direct observation, and participant
observation methods of research.

Participant observation requires that an observer become a member of his or her subjects'
community. An advantage of this method of research is the opportunity it provides to study
what actually occurs within a community, and then consider that information within the
political, economic, social, and religious systems of that community. Crosscultural research
demonstrates that Western cultural standards do not necessarily apply to other societies. What
may be normal or acceptable for one group may be abnormal or unacceptable for

Research with existing data, or secondary analysis

Some sociologists conduct research by using data that other social scientists have already collected. The
use of publicly accessible information is known as secondary analysis, and is most common in situations
in which collecting new data is impractical or unnecessary. Sociologists may obtain statistical data for
analysis from businesses, academic institutions, and governmental agencies, to name only a few sources.
Or they may use historical or library information to generate their hypotheses.

Social research is a major aspect of work for sociologists. The four main methods of research
include survey, interview, experiment, observation, and secondary analysis. Depending on
the research at hand, the choice of research methods have their pros and cons. additionally,
most research projects incorporate multiple methods.
One of the most common methods used for social research is direct analysis. Surveys and
interviews go hand in hand. Using surveys gets a broader scope of results in terms of sample
size. With technology you can submit a survey to thousands of people via the Internet using
social media or email while saving time and money compared to mailing out paper surveys.
However, with surveys you struggle with having questions that are most applicable, as well
as with getting responses to surveys.
Interviews are a micro version of surveys. You can conduct interviews by telephone, over the
Internet or in person. However, the time it takes to conduct interviews and to find willing
respondents applicable to your sample makes this a less effective measure for most research
studies. Yet if you are doing a small-scale study of a limited population, such as the graduate
students in anthropology studying at Yale in 2004, interviews are the better option.
A social experiment is the more scientific of all of the research methods. It involves having a
sociologist creating an experiment using scientific aspects with the exception that the main
subjects are humans. If a social experiment is conducted successfully it can be one of the
most reliable research methods for a social study.

There are two methods of observation: participant and nonparticipant. With participant
observation, the sociologist becomes a part of the social group being studied in order to get a
true understanding of the participants. On the other end of the spectrum, nonparticipant
observation involves a sociologist taking a step back and merely watching the participants in
There are complications with both methods. For instance, the Hawthorne Effect proves that
under conditions of being watched for sociological analysis, humans tend to change their
routines. On the other hand, if a sociologist goes undercover to watch people in a social
experiment, they have to follow protective measures, such as making sure to do so in a public

Secondary Analysis
Most studies involve secondary analysis as a way to determine what research has already
been conducted in the area of the research. Literature, sociological studies and newspaper
articles are some of the methods used for secondary analysis. Additionally, a sociologist can
listen or watch to recordings from subjects, as well as to social experiments previously
conducted. Survey data taken from previous research studies is also often used for supporting
ones upcoming research project.
Doing secondary analysis can save sociologists time and money, as they can see what has
already been studied successfully and otherwise before investing their own efforts in studies
on the same subjects.

Inductive and Deductive method

A scientific method or process is considered fundamental to the scientific investigation and
acquisition of new knowledge based upon verifiable evidence. In addition to employing the
scientific method in their research, sociologists explore the social world with several different
purposes in mind. Like the physical sciences (i.e., chemistry, physics, etc.), sociologists can
be and often are interested in predicting outcomes given knowledge of the variables and
relationships involved. This approach to doing science is often termed positivism (though
perhaps more accurately should be called empiricism). The positivist approach to social
science seeks to explain and predict social phenomena, often employing a quantitative
approach where aspects of social life are assigned numerical codes and subjected to in-depth
analyses to uncover trends often missed by a casual observer. This approach most often
makes use of deductive reasoning, which initially forms a theory and hypothesis, which are
then subjected to empirical testing.

Unlike the physical sciences, sociology (and other social sciences, like anthropology) also
often seek simply to understand social phenomena. Max Weber labelled this approach
Verstehen, which is German for understanding. This approach, called qualitative sociology,
aims to understand a culture or phenomenon on its own terms rather than trying to develop a
theory that allows for prediction. Qualitative sociologists more frequently use inductive
reasoning where an investigator will take time to make repeated observations of the
phenomena under study, with the hope of coming to a thorough and grounded understanding
of what is really going on.
Both approaches employ a scientific method as they make observations and gather data,
propose hypotheses, and test or refine their hypotheses in the formulation of theories.
The essential elements of a scientific method are iterations and recursions of the following
four steps:
1. Characterization (operationalization or quantification, observation and / or
2. Hypothesis (a theoretical, hypothetical explanation of the observations and / or
3. Prediction (logical deduction from the hypothesis or logical induction from the data)
4. Testing (informing the validity of the hypothesis by comparing it against carefully
gathered, meaningful sensory input)
A scientific method depends upon a careful characterization of the subject of the
investigation. While seeking the pertinent properties of the subject, this careful thought may
also entail some definitions and observations; the observation often demands careful
categorization, measurement and/or counting.

Hypothesis Development
A hypothesis includes a suggested explanation of the subject. In quantitative work, it will
generally provide a causal explanation or propose some association between two variables. If
the hypothesis is a causal explanation, it will involve at least one dependent variable and
one independent variable. In qualitative work, hypotheses generally involve potential
assumptions built into existing causal statements, which may be examined in a natural
Variables are measurable phenomena whose values or qualities can change (e.g., class status
can range from lower- to upper-class). A dependent variable is a variable whose values or
qualities are presumed to change as a result of the independent variable. In other words, the
value or quality of a dependent variable depends on the value of the independent variable. Of
course, this assumes that there is an actual relationship between the two variables. If there is
no relationship, then the value or quality of the dependent variable does not depend on the
value of the independent variable. An independent variable is a variable whose value or

quality is manipulated by the experimenter (or, in the case of non-experimental analysis,

changes in the society and is measured or observed systematically). Perhaps an example will
help clarify. In a study of the influence of gender (as a value) on promotion, the independent
variable would be gender/sex. Promotion would be the dependent variable. Change in
promotion is hypothesized to be dependent on gender. Similarly, in a study of gender's (as a
quality) relation to promotion, the independent variables are gender/sex and promotion, and
the dependent variable is the way people use, discuss, and / or make sense of both sex/gender
and promotion in their ongoing activities or narratives.

A useful quantitative hypothesis will enable predictions, by deductive reasoning that can be
experimentally assessed. If results contradict the predictions, then the hypothesis under
examination is incorrect or incomplete and requires either revision or abandonment. If results
confirm the predictions, then the hypothesis might be correct but is still subject to further
testing. Predictions refer to experimental designs with a currently unknown outcome. A
prediction (of an unknown) differs from a consequence (which can already be known). On
the other hand, a useful qualitative hypothesis will enable question or critique, by inductive
reasoning, of existing and / or taken-for-granted beliefs, assumptions, and theories developed
within or beyond scientific settings.

Once a prediction is made, a method is designed to test or critique it. The investigator may
seek either confirmation or falsification of the hypothesis, and refinement or understanding of
the data. Though a variety of methods are used by both natural and social scientists,
laboratory experiments remain one of the most respected methods by which to test
Scientists assume an attitude of openness and accountability on the part of those conducting
an experiment. Detailed record keeping is essential, to aid in recording and reporting on the
experimental results, and providing evidence of the effectiveness and integrity of the
procedure. They will also assist in reproducing the experimental results.
The experiment's integrity should be ascertained by the introduction of a control or by
observation of existing controls in natural settings. In experiments where controls are
observed rather than introduced, researchers take into account potential variables (e.g., the
demographics of the sample and researchers as well as the behaviours of both groups) that
could influence the findings without intention. On the other hand, in experiments where a
control is introduced, two virtually identical experiments are run, in only one of which the
factor being tested is varied. This serves to further isolate any causal phenomena.

Qualitative V/S Quantitative

Quantitative methods of sociological research approach social phenomena from the
perspective that they can be measured and/or quantified. For instance, social class, following

the quantitative approach, can be divided into different groups - upper-, middle-, and lowerclass - and can be measured using any of a number of variables or a combination thereof:
income, educational attainment, prestige, power, etc. Quantitative sociologists also utilize
mathematical models capable of organizing social experiences into a rational order that may
provide a necessary foundation for more in depth analyses of the natural world (importantly,
this element of quantitative research often provides the initial or potential insights that guide
much theoretical and qualitative analyses of patterns observed - numerically or otherwise beyond the confines of mathematical models). Quantitative sociologists tend to use specific
methods of data collection and hypothesis testing, including: experimental
designs, surveys, secondary data analysis, and statistical analysis. Further, quantitative
sociologists typically believe in the possibility of scientifically demonstrating causation, and
typically utilize analytic deduction (e.g., explore existing findings and deduce potential
hypotheses that may be tested in new data). Finally, quantitative sociologists generally
attempt to utilize mathematical realities (e.g., existing assumptions and rules embedded
within statistical practices) to make sense of natural (e.g., the experience of the actual worlds
of people) realities.
Qualitative methods of sociological research tend to approach social phenomena from
the Verstehen perspective. Rather than attempting to measure or quantify reality via
mathematical rules, qualitative sociologists explore variation in the natural world people may
see, touch, and experience during their lives. As such, these methods are primarily used to (a)
develop a deeper understanding of a particular phenomenon, (b) explore the accuracy or
inaccuracy of mathematical models in the world people experience, (c) critique and question
the existing assumptions and beliefs of both scientists and other social beings, and (d) refine
measurements and controls used by quantitative scientists via insights gleaned from the
experiences of actual people. While qualitative methods may be used to propose or explore
relationships between variables, these studies typically focus on explicating the realities
people experience that lie at the heart or foundation of such relationships rather than focusing
on the relationships themselves. Qualitatively oriented sociologists tend to employ different
methods of data collection and analysis, including: participant observation, interviews, focus
groups, content analysis, visual sociology, and historical comparison. Further, qualitative
sociologists typically reject measurement or quantities (essential to quantitative approaches)
and the notion or belief in causality (e.g., qualitative sociologists generally argue that since
there is no demonstrated possibility of ever exploring all potential variables or influences in
one study, causality is always incomplete and beyond empirical means). Finally, qualitative
sociologists generally attempt to utilize natural realities (e.g., the experience of the actual
worlds of people) to make sense of these natural realities and complicate mathematical
assumptions and rules that may lead scientists into misguided findings that lack applicability
to the actual worlds people inhabit.

Objective vs. Critical vs. Subjective

Sociologists, like all humans, have values, beliefs, and even pre-conceived notions of what
they might find in doing their research. Because sociologists are not immune to the desire to
change the world, two approaches to sociological investigation have emerged. By far the
most common is the objective approach advocated by Max Weber. Weber recognized that
social scientists have opinions, but argued against the expression of non-professional or nonscientific opinions in the classroom.[3] Weber took this position for several reasons, but the
primary one outlined in his discussion of Science as Vocation is that he believed it is not right
for a person in a position of authority (a professor) to force his/her students to accept his/her
opinions in order for them to pass the class. Weber did argue that it was acceptable for social
scientists to express their opinions outside of the classroom and advocated for social
scientists to be involved in politics and other social activism. The objective approach to social
science remains popular in sociological research and refereed journals because it refuses to
engage social issues at the level of opinions and instead focuses intently on data and theories.
The objective approach is contrasted with the critical approach, which has its roots in Karl
Marx's work on economic structures. Anyone familiar with Marxist theory will recognize that
Marx went beyond describing society to advocating for change. Marx disliked capitalism and
his analysis of that economic system included the call for change. This approach to sociology
is often referred to today as critical sociology (see also action research). Some sociological
journals focus on critical sociology and some sociological approaches are inherently critical
(e.g., feminism, black feminist thought).
Building on these early insights, the rise of Feminist methods and theories in the 1970's
ushered in an ongoing debate concerning critical versus objective realities. Drawing on early
Feminist writings by social advocates including but not limited to Elizabeth Cady
Stanton, Alice Paul, Ida Wells Barnett, Betty Friedan, and sociological theorists including but
not limited to Dorothy Smith, Joan Acker, and Patricia Yancey Martin, Feminist sociologists
critiqued "objective" traditions as unrealistic and unscientific in practice. Specifically, they along with critical theorists like Michel Foucault, bell hooks, and Patricia Hill Collins argued that since all science was conducted and all data was interpreted by human beings and
all human beings have beliefs, values, and biases that they are often unaware of and that
shape their perception of reality (see The Social Construction of Reality), objectivity only
existed within the beliefs and values of the people that claimed it. Stated another way, since
human beings are responsible for scientific knowledge despite the fact that human beings
cannot be aware of all the potential biases, beliefs, and values they use to do their science,
select their topics, construct measurements, and interpret data, "objective" or "value free"
science are not possible. Rather, these theorists argued that the "personal is political" (e.g.,
our personal decisions - no matter how small - are ultimately influenced by the political
context of our lives and ultimately will shape the personal and political realities of others
whether or not we are aware of these consequences). As a result, every scientist - regardless
of their intentions and / or awareness - may seek to follow Weber's advice concerning
objective teaching and research, but must also remain aware that they will ultimately fail to

achieve this ideal. Whether or not scientists explicitly invoke their personal opinions in their
teaching and research, every decision scientists make will ultimately rely upon - and thus
demonstrate to varying degrees - their subjective realities. As a result, current debates
typically center around objective (an ideal) versus subjective (data-based) interpretations of
science while scholars continue to debate the merits and limitations of subjective/objective
versus critical approaches.
Some examples of the subjective basis of both "objective" and "critical" sociology may
illustrate the point. First, we may examine the research process for both objective and critical
sociologists while paying attention to the many decisions people must make to engage in any
study from either perspective. These decisions include:

The selection of a research topic (this selection reveals something the author believes is
important whether or not it is)

The selection of data (this selection reveals data the author believes is reliable whether or
not it is)

If the researcher decides to collect zir own data, then they must:

Decide where to collect data

Decide who to collect data from

Decide what questions to ask (which ones they believe will answer the question) and
how to ask these questions (which forms of talk they believe are best for getting the
answers they want)

Decide how much data to collect

Decide how to analyse the data collected (if mathematically, which protocols will be
used and which software program, and if qualitatively which themes will ze look for
and / or what software program)

Decide how to measure or categorize the data (if mathematically, what set of
parameters counts as a good measure, and if qualitatively what must a category

Decide how to interpret the measurements or categories (if mathematically, what

exactly do the numbers mean socially, and if qualitatively what do the categories say
about society)

Decide how to discuss the interpretation (which theories should be used and which
ones should be ignored)

If the researcher decides to use secondary data, this becomes even more complicated.
While they will have to do the final four items listed above, they must also:

Trust that the data collection occurred properly

Trust that the data was organized properly

Trust that the questions were answered properly

Trust that the sample is appropriate

As you can see above, the research process itself is full of decisions that each researcher must
make. As a result, researchers themselves have no opportunity to conduct objective studies
because doing research requires them to use their personal experiences and opinions (whether
these arise from personal life, the advice of the people that taught them research methods, or
the books they have read that were ultimately subject to the same subjective processes)
throughout the process. As a result, researchers can - as Feminists have long argued - attempt
to be as objective as possible, but never actually hope to reach objectivity. This same problem
arises in Weber's initial description of teaching. For someone to teach any course, for
example, they must make a series of decisions including but not limited to:

Deciding what subjects to cover within the overall course

Deciding which readings to use to convey information

Deciding what measures of learning will be used and what measures will be left out of
the course

Deciding what counts as an appropriate or inappropriate answer on any and all measures
used in the course

As a result, Weber's objectivity dissolves before the teacher ever enters the classroom.
Whether or not the teacher (or researcher) explicitly takes a political, religious, or social
stance, he or she will ultimately demonstrate personal stances, beliefs, values, and biases
implicitly throughout the course.
Although the recognition of all science as ultimately subjective to varying degrees is fairly
well established at this point, the question of whether or not scientists should embrace this
subjectivity remains an open one (e.g., to be or not to be political in classrooms and research
projects). Further, there are many scientists (in sociology and other sciences) that still cling to
beliefs about objectivity, and thus promote this belief (political in and of itself) in their
teaching, research, and peer review. As a result, the debate within the field continues without
resolution, and will likely be an important part of scientific knowledge and scholarship for
some time to come.

Correlation and Causation

In the scientific pursuit of quantitative prediction and explanation, two relationships between
variables are often confused: correlation and causation. While these terms are rarely used in
qualitative science, they lie at the heart of quantitative methods, and thus constitute a
cornerstone of scientific practice. Correlation refers to a relationship between two (or more)
variables in which they change together. A correlation can be positive/direct or
negative/inverse. A positive correlation means that as one variable increases (e.g., ice cream
consumption) the other variable also increases (e.g., crime). A negative correlation is just the

opposite; as one variable increases (e.g., socioeconomic status), the other variable decreases
(e.g., infant mortality rates).
Causation refers to a relationship between two (or more) variables where one variable causes
the other. In order for a variable to cause another, it must meet the following three criteria:

the variables must be correlated

change in the independent variable must precede change in the dependent variable in time

it must be shown that a different (third) variable is not causing the change in the two
variables of interest (a.k.a., spurious correlation)

An example may help explain the difference. Ice cream consumption is positively correlated
with incidents of crime.
Employing the quantitative method outlined above, the reader should immediately question
this relationship and attempt to discover an explanation. It is at this point that a simple yet
noteworthy phrase should be introduced: correlation is not causation. If you look back at the
three criteria of causation above, you will notice that the relationship between ice cream
consumption and crime meets only one of the three criteria (they change together). The real
explanation of this relationship is the introduction of a third variable: temperature. Ice cream
consumption and crime increase during the summer months. Thus, while these two variables
are correlated, ice cream consumption does not cause crime or vice versa. Both variables
increase due to the increasing temperatures during the summer months.
It is important to not confound a correlation with a cause/effect relationship. It is often the
case that correlations between variables are found but the relationship turns out to be
spurious. Clearly understanding the relationship between variables is an important element of
the quantitative scientific process.

Typically a population is very large, making a census or a complete enumeration of all the values in that
population infeasible. A 'sample' thus forms a manageable subset of a population. In positivist research,
statistics derived from a sample are analysed in order to draw inferences regarding the population as a
whole. The process of collecting information from a sample is referred to as 'sampling'. Sampling methods
may be either 'random' (random sampling, systematic sampling, stratified sampling, cluster sampling) or
non-random/nonprobability (convenience sampling, purposive sampling, snowball sampling).The most
common reason for sampling is to obtain information about a population. Sampling is quicker and cheaper
than a complete census of a population.

Brief: Sociological Research Methods

General Approaches
Empirical Research
Sociological research is based on the use of empirical data to substantiate concepts and
theories and to test hypotheses.

Empirical data: Facts that we observe, measure, and verify with our senses.
Concept: A simple, abstract construct (idea) that represents some aspect of the world.
Theory: A formal statement that attempts to explain a phenomenon by attributing it to
particular relationships among a group of concepts.
Hypothesis: An educated guess or proposition about the relationship between two or more
phenomena that is stated in testable form.
Sociology: Science or Interpretation
Most sociologists probably find themselves somewhere between these two positions:
Sociology as science: Sociological research is a systematic method of direct observation of
the world, similar to the natural sciences, which produces objective knowledge of social
phenomena and, in some cases, general social laws. Associated with variable research.
Interpretive sociology: Sociological research examines the meanings that actors attach to
social phenomena. Meanings are subjective and not governed by universal laws; hence,
sociology differs from natural science. Associated with qualitative research.

Research Design
Once the researcher has a question and some concepts and theories, she or he must pick a
level of analysis, a time frame, and a method of gathering data, and decide what type of data
analysis will be most appropriate.
Level of Analysis
Unit of analysis: The specific social entity about which data will be gathered and empirical
claims made. Some possible units of analysis: individuals, careers, city birth rates,
unionization votes, nations, business establishments.
Cross-sectional study: Uses data from one time point only.
Longitudinal study: Uses data gathered at several points in time. Permits conclusions about
Methods of Gathering Data
Surveys: People are asked to respond to a prepared set of questions or statements in either a
verbal interview or a written questionnaire.
In-depth interviews: People are asked to respond at length to a series of questions posed by
the researcher. Questions may be fixed in advance or the interviewer may allow open-ended
Field research (participant observation): Researchers observe and talk to people in their
ordinary settings while sometimes joining in their activities.
Document study: Data is gathered from documents such as newspaper articles, marriage
records, or diaries.
Experiments: A method used to test a specific hypothesis about a cause and effect
relationship. An experiment has three steps: (1) measuring the effect variable; (2) exposing
the effect variable to the cause variable; and (3) measuring the effect variable again to see if a
change has occurred. Any factors that might affect the two variables being measured and that
are not part of the causal relationship being tested must be controlled.
Variable Analysis or Qualitative Analysis?
Variable research: Entails choosing variables to represent relevant concepts, measuring the
variables, and analysing the results. Data is often gathered through surveys; analysis is
Variable: A concept that can take on more than one value. For example, the variable
ethnicity may take on the values African-American, Latino, Asian, etc.


Measurement: The procedure by which the value of a variable is determined in a specific

case. For example, one could measure ethnicity by looking at each individual person or by
asking each person what their ethnic identification is.
Qualitative research: Entails selection of questions, concepts, and relevant data sources.
Data is often gathered through interviews or field research. Analysis involves identification
of categories and patterns in the data and continual reassessment of questions and concepts.
Causal Analysis: A Dominant Approach
Much sociological research aims at establishing the presence of causal relationships among
social phenomena. (The following discussion assumes variable research.)
Causal analysis: The goal is to establish (or refute) the existence of a causal relationship
between two or more variables. To establish causation, the research must demonstrate that:
(1) the variables are correlated; (2) the causal variable precedes the effect variable in time;
and (3) a change in the causal variable results in a change in the effect variable regardless of
changes in other factors. Proving (3) is difficult because of the broad potential for
unmeasured spurious or intervening relationships in the social world.
Correlation: Two variables are correlated if they change together.
Positive correlation: When the value of one increases (decreases), the value of the other
increases (decreases).
Negative correlation: When the value of one increases (decreases), the other decreases
Spurious relationship: A false relationship between two variables (A, B). A and B may
appear to be causally related, but they are actually affected independently by a third variable
(C). For example, suppose that the U.S. cities with the highest number of art museums (A)
also have the highest concentrations of smog (B). Does this mean that art causes smog? A
more likely explanation is that city size (C) is causally related to both the number of
museums and the concentration of smog. A C B
Intervening relationship: A relationship between two variables (A, B) that is dependent on
the actions of a third variable (C). For example, suppose that working-class students (A)
perform poorly on SATs (B). Does this mean that working-class students are less intelligent?
A more likely explanation is that working-class students go to low-quality schools (C). Here,
a school-quality variable intervenes in the relationship between student social class and SAT
score. A C B
Issues in Approach
The following concerns are relevant to all sociological research:
Reliability: Consistency of observation, such that the same results are obtained each time the
observation is repeated.
Validity: There are many types of validity; one important type is construct validity, which
addresses the question of whether the researcher is measuring exactly what he or she claims
to measure.
Generalizability: Most sociologists can only observe a few of the sociological phenomena
about which they wish to make empirical claims, so they generalize from this few to the
larger group. The most common way of doing this is to select the smaller group of cases by
systematically sampling them from the larger group or population. If the sample represents
the population well, conclusions about it are generalizable; if not, they are biased. Typically,
the best way to achieve generalize ability is by using a large, randomly selected sample
Value-free research and objectivity: Many believe that sociologists should strive to
produce objective, value-free analysis; they should avoid introducing their own values into
their research. Others claim that it is impossible and undesirable for researchers to completely
suppress their values; values are an important impetus for sociological research.

Research ethics: Sociologists agree that some research methods have the potential to harm
or pose risks to participants. For this reason, most sociologists adhere to established
guidelines for conducting research in a manner that will reduce risks and conform to widely
accepted ethical standards.