You are on page 1of 51

RESEARCH METHODS

CHAPTER 2

SOCIAL PSYCHOLOGY IS
A SCIENCE
Findings in social psychology are based on the
scientific method. Scientific research describes a
phenomenon (are people more aggressive when it
is very hot outside?), makes predictions about it
(will the rate of homicide increase during the
summer?), and explains why it happens (why do
people act more aggressively when it is hot?).

STEPS IN THE RESEARCH PROCESS


Step 1
Form a
Question
Step 6
Propose or
Revise a
Theory

Step 2
Search the
Literature

Step 5
Collect and
Analyze
Data
Step 4
Create an
Operational
Definition

Step 3
Form a
Hypothesis

FORM A QUESTION
All research in social psychology and other
scientific fields starts with a question. Many studies
start with a question based on observation of a
real-world event.
Sometimes people form questions based on their
own experiences of observations.
Sometimes these questions are designed to test
established theories in psychology.

QUESTION
Many other hypotheses come from observations of
everyday life, such as why people failed to help
murder victim Kitty Genovese.
Thirty-eight neighbors failed to call police during
her prolonged and violent murder.
Genoveses neighbors might have assumed that
someone else had called the police.
Why did people fail to call police?

SEARCH THE LITERATURE


Because some ideas that you have are likely to
have been studied by others, it is essential to start
investigating these ideas by figuring out what other
people might have found about the same, or
similar, ideas.
This can be conducted by going to the library,
reading journal articles, or online internet searches.

FORM A HYPOTHESIS
A hypothesis is a testable prediction about the
conditions under which an event will occur. It is a
statement about the expected cause and effect
relationship between two variables, but has greater
specificity than the original question you asked.
Because a hypothesis can be directly tested, it
includes a specific prediction.

HYPOTHESIS
Perhaps the bystanders would have been more
likely to help had each thought he or she alone was
witnessing the murder.
Latan and Darley (1968) called this diffusion
of responsibility.
Once a researcher has a hypothesis, how can he or
she tell if it is right? In science, idle speculation will
not do; the researcher must collect data to test a
hypothesis.

CREATE
AN OPERATIONAL DEFINITION
An operational definition describes a specific
procedure or measure of how you will test the
hypothesis.
Researchers can define their variables in very
different ways, which in turn can influence the
findings.

COLLECT AND
ANALYZE DATA
Data can be collected in a number of ways by
observation, surveys, or experiments.
Analyzing the data is important to determine if the
ideas are supported by the data collected.

PROPOSE AND/OR
REVISE A THEORY
The final step in the research process is proposing a theory,
which is an organized set of principles used to explain
observed phenomena. Although hypotheses are specific
predictions about the association between two events, they
do not explain how or why these two events are connected.
In contrast theories provide potential explanations.
Sometimes results from a study can lead to the revision of a
particular theory or even refute a theory.
Because theories provide one type of explanation for a given
phenomena, they also generate questions for future
research.

CORRELATIONAL RESEARCH METHODS


Correlation is a technique that examines the extent
to which two or more variables are associated with
one another (e.g. height and weight are highly
correlated).
The conditions on how questions are asked can
influence results.
Direct vs. indirect approach

OBSERVATIONAL OR
NATURALISTIC METHODS
Observational or naturalistic methods are used to
describe and measure peoples behavior in everyday
situations. In this approach researchers observe
behavior and systematically record that behavior.
Emile Durkheim Sociologist who conducted a naturalistic
study on suicide in the mid-1800s.

Some researchers collect naturalistic data without


directly observing peoples behavior.
Differing alcohol consumption at various university housing
situations.

ARCHIVAL RESEARCH
META-ANALYSIS
Archival research is an approach that uses existing
recorded behavior, i.e. divorce rates, death rates,
crime rates, etc.
Meta-analysis is a literature review that analyzes
data from many different studies and sources. It
uses a statistical technique for combining data that
have been collected by different researchers.

ADVANTAGES OF
OBSERVATIONAL METHODS
Since observational and naturalistic methods are
based in the observation of real-world phenomena, the
help researchers develop hypotheses and theories.
These methods are also relatively easy to conduct.
Naturalistic methods can provide large amounts of
data that researchers would never be able to collect on
their own.
This is particularly important when researchers are
interested in examining how something has changed
over time.

LIMITATIONS OF
OBSERVATIONAL METHODS
1) The presence of the observer can influence behavior.
People are likely to modify their behavior when they are being
monitored.

2) The observers own biases can influence how they


interpret the behavior observed.
Inter-rater reliability: Researchers often have at least two
people do the ratings independently and then measure how
often they agree.

3) Correlation vs. Causation


Observational methods can show whether two variables are
associated (correlated), with each other, but they cannot tell
us what causes the other. There could be a third variable that
explains the observed association between the two variables.

LIMITS OF CORRELATION METHOD


If two variables (e.g., TV violence & aggression) are
correlated, there are three possible causal
relationships:
Maybe TV violent makes the viewer become
violent.
Maybe kids who are already violent are more
likely to watch violent TV.
Maybe both are caused by something else like
parental neglect.

Correlation coefficients are expressed as numbers


that can range from 1.00 to +1.00.
1.00 means that two variables are perfectly
correlated in a positive direction.
0 means that two variables are not correlated
1.00 means that two variables are perfectly
correlated in a negative direction.
In everyday life, of course, perfect correlations are
rare.

POSITIVE CORRELATION
Increases in the value of one variable are
associated with increases in the value of the other
variable.
Students who receive higher grades in a class
give that professor more positive teaching
evaluations (positive correlation).
We still dont know what causes the other.
Does having a better professor cause the students to get
better grades?
Do students doing well in the class come to like their
professor more?

NEGATIVE CORRELATION
Increases in the value of one variable are
associated with decreases in the value of the other
variable.
Vaccination rate correlates negatively with
disease rate: The more often people get
vaccinated, the less often people get the disease.
People are hostile have fewer friends. However,
you cant tell if people who are mean to others
have trouble making friends, of if people who
dont have many friends grow to be hostile over
time.

CORRELATION VS. CAUSATION


TOBACCO
Hill and Doll (1950s Tobacco case smoking and
lung cancer) show us that correlation should not be
dismissed too easily.
However, they also show we shouldnt give up on
the search for causal explanations. They
painstakingly continued their research, and
evidence of a causal association soon mounted.

CORRELATION VS. CAUSATION


TOBACCO
Hill and Doll took a pragmatic approach in their search for
causation.
Is there a dose-response relationship?
Yes: Heavy smokers are more likely to suffer from lung cancer.

Does the timing make sense?


Yes: Smokers develop cancer long after they begin to smoke. This
contradicts Fishers alternative hypothesis that people self-medicate
with cigarettes in the early stages of lung cancer.

Do multiple sources of evidence add up to a coherent picture?


Yes: When doctors heard about what Hill and Doll were finding, many of
them quit smoking, and it became possible to see that the quitters were
at lower risk of lung cancer.

Conclusion: We should respect correlation, but it is a clue to a


deeper truth, not the end of our investigations.

SELF-REPORT OR
SURVEY METHODS
Surveys rely on asking people questions about
their thoughts, feelings, desires, and actions.
To make sure that the results are generalizable,
researchers randomly select survey respondents
from the population at large.
Event-recording: Respondents report various
experiences they have at regular intervals.

HOW DO YOU FEEL


ABOUT YOURSELF?
Self-Esteem Scale
Instructions: Rate your agreement with each of
these items on a 1 to 4 scale, with 1 meaning
strongly disagree and 4 meaning strongly agree.

1. I feel that Im a person of worth, at least


on an equal plane with others.
2. On the whole, I am satisfied with myself.
3. I wish I could have more respect for
myself.
4. I certainly feel useless at times.
5. At times I think I am no good at all.
6. I feel that I have a number of good
qualities.
7. All in all, I am inclined to feel that I am a
failure.
8. I am able to do things as well as most
other people.
9. I feel that I do not have much to be
proud of.
10.I take a positive attitude toward myself.

Scoring: Indicate whether you


strongly agree, agree, disagree,
or strongly disagree with each of
these statement. For statements
1, 2, 6, 8, and 10 give yourself 4
points for strongly agree, 3
points for agree, 2 points for
disagree, 1 point for strongly
disagree. For statements 3, 4, 5,
7, and 9 give yourself 1 point for
strongly agree, 2 points for
agree, 3 points for disagree, and
4 points for strongly disagree.
Then add up your total number
of points.

Interpretation: People with


a higher score on this scale
have higher self-esteem.

ADVANTAGES OF SURVEYS
Surveys can collect information about the link
between peoples attitudes and behaviors.
Researchers can judge the relationship between
variables that are difficult to observe, such as
how often people engage in safe sex.
The ability to sample representative segments of
the population.
As long as the sample is selected randomly, we
can assume that the responses are a reasonable
match to those of the population as a whole.

LIMITATIONS OF SURVEYS
Some factors that limit the reliability of self-reports
and surveys:
1) Question wording:
2) Response options:
3) Inaccuracy of responses:

QUESTION WORDING
1) Biased findings as a result of leading questions.
Leading questions are those questions that provide some evidence
of the right answer based on how they are phrased.
How much are you in favor of allowing an abortion for a teenager who
becomes pregnant following a rape?
How much are you in favor of allowing a woman to murder an innocent
baby in her womb?

2) Subtle wording differences can lead to different results and


the order in which questions are asked can influence the
response.
Preceding questions influence our interpretation of following questions.
Asking how happy people are with their life followed by their marriage gives
differing answers if asked in reverse.

3) Providing information about who is conducting the research


influences responses.

RESPONSE OPTIONS
Options given in surveys can influence the results.
Responses provide people an idea of what the normal
or typical behavior is, and people dont want to
appear different from others, and especially not worse
than others.
Therefore, they are likely to choose one of the mid-level choices
as opposed to one of the extreme options.
E.g.: Daily cigarette consumption

Responses can have an even stronger impact on


answers when participants must choose between a set
of very limited options.
Respondents may choose options that are closer to their true
feelings since there was no option to select/express their true
feelings.

INACCURACY IN RESPONSES
Survey methods are limited by the possibility of
inaccurate reporting.
People might believe they are telling the truth,
but may simply not be able to remember the
necessary information.
Event-recording measures can reduce this problem.

Why do people provide inaccurate information?


People are concerned about the social desirability of their
answers, especially in cases involving highly personal or
controversial topics, i.e. sexual issues (condom use,
abortions, pornography).

COVERT MEASURES
Covert measures are not directly under a persons
control, which helps to minimize the problems
associated with socially desirable responding.
Can be used in which participants might not want
to be honest in their responses.
Sensitive topics: sexual (homosexuality), prejudice
(stereotypes, discrimination), aggression (revenge).

EXPERIMENTAL METHODS
In experimental methods researchers manipulate one
or more independent variables and then measure the
effects on one or more dependent variables.
Independent variable: The variable that is manipulated in
experimental research.
Dependent variable: The factor that is measured to see if it is
affected by the independent variable.

This approach determines systemically whether the


independent variable caused the dependent variable,
and therefore provides evidence of causation as
opposed to correlation.

AN EXPERIMENTAL METHOD
Researchers conducted an experiment to find out if
women eat less when they want to appear attractive.
1st condition: women were informed they were interacting with a very
desirable (independent variable) man (athletic, handsome,
educated, artistic, etc.)
2nd condition: women were informed they were interacting with a less
desirable (independent variable) man (boring, stays at home, only
interested in money, etc.)

Researchers measured the amount of candy


(dependent variable) women eat in both conditions.
As predicted, women ate less in the first condition.
Men ate about the same regardless of whether the
partner was attractive.

INDEPENDENT & DEPENDENT


VARIABLES
Latan and Darley (1970)
Independent variable: Number of people supposedly
present when a researcher pretends to have a seizure.
Dependent variable: Number of people who try to help in
the emergency.
Outcome:
When participants believed four other people witnessed
the seizure, only 31% offered assistance.
When participants believed only two others witnessed
the seizure, helping behavior increased to 62%.
When each participant believed that he or she was the
only witness, nearly everyone helped (85%).

RANDOM ASSIGNMENT
Random assignment is a technique of ensuring that
a sample of people is representative of a
population by giving everyone in the population an
equal chance of being selected for the sample.
Flipping a coin, drawing numbers, etc.
This technique allows for greater confidence in the
research findings.

CONTROL
Researchers have a lot of control over what
happens to participants in their experiment. You
can choose what happens to whom, when, and
how.

INTERNAL VALIDITY
Internal validity is making sure that nothing besides
the independent variable can affect the dependent
variable.
This is accomplished by controlling all extraneous
variables, reducing or eliminating the demand
characteristics (cues in a research setting that may
guide participants behavior), and by randomly
assigning people to different experimental conditions.
1)
2)
3)
4)

Provide a good cover story.


Provide a high quality control condition.
Minimize experimenter expectancy effects.
Design studies with high experimental realism.

EXTERNAL VALIDITY
For all the advantages of the experimental
method, there are some drawbacks.
By virtue of gaining enough control over the
situation so as to randomly assign people to
conditions and rule out the effects of extraneous
variables, the situation can become somewhat
artificial and distant from real life.
External validity is the extent to which the
results of a study can be generalized to other
situations and to other people.

INCREASING
EXTERNAL VALIDITY
1) Mundane realism:
The extent to which an experiment is similar to real-life
situations.
Conducting field experiments in a natural setting.

2) Use random/representative sample:


A sample that reflects the characteristics of the population
at large.
Try to avoid convenience sampling.

3) Make participation convenient:


4) Conduct Replications:
Conduct the same study in different populations or
locations.

REVIEWING VALIDITY
Even though Latan and Darley (1970) staged an
emergency that in significant ways was unlike ones
encountered in everyday life . . .
Was it psychologically similar to real-life emergencies?
Were the same psychological processes triggered?
Did the participants have the same types of
perceptions, thoughts, decisions, and behaviors that
they would in a real-life situation?
If so, then the study is high in psychological realism
and we can generalize the results to everyday life.

KEY POINTS
The only way to be certain an experiments
results represent the behavior of a particular
population is to ensure that the participants are
randomly selected from that population.
Unfortunately, it is impractical and expensive to
select random samples for most social
psychology experiments.
Many researchers address this by studying basic
psychological processes so fundamental that they
are presumably universally shared.

REPLICATION
The question then is, how can researchers tell
whether the processes they are studying are
universal?
How can we trust that a study done with only
college sophomores captures everyday
responses?
The ultimate test of an experiments external
validity is replication.

SOLUTION
Several studies might
find an effect of the
number of bystanders
on helping behavior, for
example, while a few do
not.
How can we make
sense of this?

Meta-Analysis:
A statistical technique
that averages the
results of two or more
studies to see if the
effect of an
independent variable is
reliable.

WHAT IS THE BEST APPROACH?


There is no single best method, all methods have their
strengths and weaknesses.
Experiments randomly assign people to conditions
that allows for examining whether a specific factor is
likely to cause another (causation).
However, experiments are somewhat artificial and cant
accurately provide much information in real-life situations.

Naturalistic observation methods gives us accurate


information about what happens in the real world.
However, naturalistic observation tells us more about how two
or more variables are connected (correlated).

In sum, different methods are best for providing different


types of information and for answering different
questions.

THE BASIC DILEMMA OF


THE SOCIAL PSYCHOLOGIST
One of the best ways to increase external validity is
by conducting field experiments.
In a field experiment, researchers study behavior
outside of the laboratory, in its natural setting.
A field experiment has the same design as a
laboratory experiment except that it is conducted in a
real-life setting (sidewalk, store, street, campus
grounds).
Participants in a field experiment are unaware that the
events they experience are in fact an experiment.
External validity of such an experiment is high, since,
it is taking place in the real world with real people.

THE BASIC DILEMMA OF


THE SOCIAL PSYCHOLOGIST
There is almost always a trade-off between internal and
external validity in social psychological research.
By increasing internal validity, some external validity
(generalizability) is sacrificed.
By increasing external validity (e.g., by conducting a
field experiment), researchers often lose control over
the setting and sacrifice internal validity.
Researchers often begin by maximizing internal
validity, so that they know what is causing what, and
then establishing external validity with replications in
different settings and with different populations.

THE BASIC DILEMMA OF


THE SOCIAL PSYCHOLOGIST
The way to resolve this dilemma is not to try to do
it all in a single experiment.
Most social psychologists opt first for internal
validity, conducting laboratory experiments.
Other social psychologists prefer to maximize
external validity by conducting field studies.
Many social psychologists do both.
Through replication, a given research question
can thus be studied with maximum internal and
external validity.

ETHICAL ISSUES
To create realistic, engaging situations, social
psychologists frequently face an ethical dilemma:
For scientific reasons, we want our experiments to
resemble the real world as much as possible and to
be as sound and well controlled as we can make
them.
But we also want to avoid causing our participants
undue and unnecessary stress, discomfort, or
unpleasantness.
These two goals often conflict as the researcher goes
about the business of creating and conducting
experiments.

ETHICAL ISSUES
Social psychologists are concerned with the welfare
of their research participants.
Example: Zimbardo (1973) prison study.
Researchers also make discoveries that can benefit
society.
To gain insight into such critical issues, researchers
must create vivid events that are involving for the
participants.

GUIDELINES FOR
ETHICAL RESEARCH
1. Review by an Institutional Review Board.
2. Provide Informed Consent.
Agreement to participate in an experiment, granted in full
awareness of the nature of the experiment, which has been
explained in advance.
Deception maybe involved depending on the study, i.e.
not providing participants with accurate information.

3. Protect Confidentiality.
4. Provide Debriefing.
Debriefing participants afterwards about the purpose of the
study and what transpired, especially if there was any
deception involved.

CULTURAL INFLUENCE ON
RESEARCH
Researchers always have to guard against
imposing their own viewpoints and definitions,
learned from their culture, onto another culture
with which they are unfamiliar.
They must also be sure that their independent and
dependent variables are understood in the same
way in different cultures.

CULTURAL INFLUENCE ON
RESEARCH
1. The impact of question ordering.
How questions are ordered can influence the results of a
study.

2. The impact of question wording.


How questions are worded in surveys can influence
responses in different ways across cultures.

3. The impact of language.


How does language affect the meanings of words.

You might also like