You are on page 1of 46

Chapter 2

Doing Social Psychology


Research

Copyright © 2017 Cengage Learning. All Rights Reserved.


Why Should You Learn About 

Research Methods?

We are bombarded with information in our daily lives, like the countless
advertisements designed to persuade us to buy particular products or
adopt particular opinions!
Learning the methods used in social psychology research can help you
become a more sophisticated consumer of information and weed out fact
from fiction.
Why Should You Learn About 

Research Methods?

• You can become a better, more sophisticated information consumer


• You can improve your reasoning about real-life events and problems
– We’re constantly bombarded with “facts” from the media, sales pitches, and
from other people!
– Much of this information is oversimplified, misleading, or just plain wrong!
– Ex: we are told about the health benefits of eating certain foods, benefits of
exam prep courses, social status of driving a certain car or wearing certain
clothes/shoes
– When we see these things in our daily life and/or on social media, we SHOULD
be saying, “Prove it! What is the evidence?” What alternative explanations
might there be?”
• You’ll learn to think like a scientist
– Critical thinking!
– Healthy sense of doubt about claims you see/read
– Ability to critically evaluate information you are exposed to in your daily life
and separate the fact from the fiction
• Bonus: you may perform better on tests and in future courses!
Copyright © 2017 Cengage Learning. All Rights Reserved.
Developing Ideas

Beginning the Research Process

Copyright © 2017 Cengage Learning. All Rights Reserved.


Getting Ideas and Finding Out What’s
Been Done

• The research process involves:


– Coming up with ideas
– Refining them
– Testing them
– Interpreting the meaning of the results obtained
• Let’s focus on the first stage, coming up with ideas:
– First, start asking questions
• These questions can come from a variety of sources
• Ex: from something perplexing like why are men less likely to ask for help than women
• Ex: from something amusing, like song lyrics suggesting patrons in a bar seem more attractive as
closing time approaches
• Each of these questions can be studied scientifically!
– Search the literature and get ideas from research already published and the new
questions the researchers have raised for future study
• Textbooks
• Internet databases
• Published articles and books

Copyright © 2017 Cengage Learning. All Rights Reserved.


Hypotheses and Theories

• From your idea, create a hypothesis:


– An explicit, testable prediction about the conditions under which
an event will occur
– Based on observation, existing theory, or previous research findings
– Ex: “teenage boys are more likely to be aggressive toward others
if they have just played a violent video game for an hour than if
they played a nonviolent video game for an hour.”
– Remember, it must be specific and able to be tested empirically!
– As data are collected to test the hypothesis, a more advanced step
in the research process takes place, the proposal of a theory.
• Theory:
– An organized set of principles used to explain observed phenomena
• Ideally, they are efficient and precise; encompass all relevant
information; and lead to new hypotheses, further research, and
better understanding

Copyright © 2017 Cengage Learning. All Rights Reserved.


Basic and Applied Research

• Basic research
– Goal is to increase our understanding of human behavior
– Often designed to test a specific hypothesis from a specific
theory
• Applied research
– Focuses more specifically on making applications to the world
and contributing to the solution of social problems
• Despite their differences, Basic & Applied research are closely
connected in social psychology
– Some studies actually test a theory and a real-world phenomenon at the same
time
– Ex: testing theories about the effects of people’s beliefs about human abilities
while also addressing important problems like students dropping out of school

Copyright © 2017 Cengage Learning. All Rights Reserved.


Refining Ideas

Defining and Measuring 



Social Psychological Variables

Copyright © 2017 Cengage Learning. All Rights Reserved.


Defining & Measuring Social
Psychology Variables

• Fact: You cannot test your hypothesis before you’ve decided how
you will define and measure the variables you’re interested in!
– This is sometimes straightforward and other times more difficult
• Ex: comparing how quickly people run a 100-meter dash when alone
and when racing against another person
– You can easily use a stopwatch to record speed and can easily
observe runners racing either alone or in pairs
• More difficult example: what if you’re interested in studying the
effects of mood on altruistic (helping) behavior?
– What do you mean by “mood”? How will you define it?
– How will you measure mood?
– How will you define “altruistic behavior?”
– You will need to clearly define these concepts and decide how
you will measure them before you begin your empirical study.

Copyright © 2017 Cengage Learning. All Rights Reserved.


Defining & Measuring Variables

• Take this picture for another example


• How would you define/measure a conceptual variable like loneliness?
– Would you observe the boy’s behavior?
– Ask how he feels?
– A combination of the two?
• These are examples of approaches to defining and measuring conceptual
variables.

Copyright © 2017 Cengage Learning. All Rights Reserved.


From the Abstract to the Specific

• Conceptual variables
– Abstract or general variables
– This is usually how the variables start when developing a hypothesis
– Ex: prejudice, conformity, attraction, love, anxiety
– In order to test your hypothesis, you must transform these conceptual
variables into very specific, measurable ones
• Operational definition
– States how the conceptual variable will be manipulated or measured
– Transforms the variable from the abstract (conceptual) to the specific
(operational)
• Construct validity
– Used to evaluate the manipulation and measurement of variables
– The extent to which the manipulations in an experiment really
manipulate the conceptual variables they were designed to and
– The extent to which the measures used really measure the conceptual
variables they were designed to.
Copyright © 2017 Cengage Learning. All Rights Reserved.
Operational definitions

• Imagine you wanted to conduct a study on the effects of


alcohol intoxication on aggression.
• One conceptual variable involved would be intoxication
– How will you define and measure it?
– One researcher might operationally define intoxication as when a
participant has a blood alcohol level of .10 or more and use a
breathalyzer to measure it
– Another researcher might define intoxication as when a
participant says that he or she feels drunk and use their verbal or
written statements to measure it
• Another conceptual variable involved is aggression
– Some ethical constraints here – you can’t let participants attack
each other. So how will you measure it?
– Some researchers have used behaviors like administering shocks
or blasts of noise to another person to measure it
Operational Definitions Continued

• There is no one single best way to transform a variable from the abstract
(conceptual) to the specific (operational).
• It is the researcher’s job to consider the advantages and disadvantages of
different approaches.
• They spend a great deal of time fine-tuning their operational definitions
to find the best way to capture the concepts they want to study.
• This image from your text shows one way of measuring height!

Copyright © 2017 Cengage Learning. All Rights Reserved.


Measuring Variables: Using Self-
Reports
• Self-reports
– One way researchers measure variables; widely used
– Participants disclose their thoughts, feelings, desires, and
actions
– Ex: the Rosenberg Self-Esteem Scale
• Consists of questions that measure overall self-esteem
• Participants indicate to what extent they agree with statements such as “I
feel that I have a number of good qualities” and “All in all, I am inclined to
feel that I’m a failure”
• This scale was found to have good construct validity (if you don’t know
what I mean by construct validity, review the last several slides before
moving on)

Copyright © 2017 Cengage Learning. All Rights Reserved.


Self-Reports Continued
• Problems with self-reports
– Not always accurate and can be misleading
• Our desire to look good to ourselves and others can influence how we respond (not
always responding honestly)
– Also affected by the way questions are asked (i.e., the wording or order)
• When condoms said to have “95% success rate,” a majority (88%) of participants
indicated they thought condoms were effective.
• When said to have “5% failure rate,” less than half (42%) indicated condoms were
effective.
• These statements are merely another way of saying the same thing, but the wording
affected responses!

• Bogus Pipeline Technique


– A procedure where participants are led to believe their responses will be verified
by an infallible lie detector
– When participants believe their lies will be detected, they report facts about
themselves more accurately and endorse socially unacceptable opinions more
frequently
– The bogus pipeline is bogus! No such device exists.
• Nonetheless, belief in its power discourages people from lying.

Copyright © 2017 Cengage Learning. All Rights Reserved.


Self-Reports Continued

• Can also be inaccurate because our memories for past thoughts or


behaviors are not always reliable!
– Our memories are very prone to error, especially if how they feel now about
something is different from how they felt in the past

• Researchers have developed strategies to minimize this problem by


reducing the time that elapses between an actual experience and
the person’s report of it.
– Interval-Contingent Self-Reports: respondents report their experiences at
regular intervals, usually once a day
– Signal-Contingent Self-Reports: respondents report their experiences as soon as
possible after being signaled to do so, usually by text message or a special app
– Event-Contingent Self-Reports: respondents report on a designated set of events
as soon as possible after such events have occurred

Copyright © 2017 Cengage Learning. All Rights Reserved.


Measuring Variables: 

Using Observations

• Another tool used to measure variables is observing people’s actions


• Observations can be simple or elaborate
• Interrater reliability: level of agreement among multiple observers of the
same behavior
– Necessary when observations are more elaborate, such as judging whether someone is
acting warmly or coldly toward another person
– Multiple observers record the same behavior and examine their agreement
– Only when different observers agree can the data be trusted (and the operational
definition)
• Advantages
– Avoids our sometimes-faulty recollections and distorted interpretations
of our own behavior; more objective
• Disadvantages
– Risks altering behavior of the observed (might influence the
respondent)
– When we know we’re being observed, might be biased by desire to
present ourselves in a favorable light
Copyright © 2017 Cengage Learning. All Rights Reserved.
Measuring Variables: Using Technology

• Various types of technology are used to


measure cognitive and physiological
responses
– Reaction time
– Heart rate
– Eye tracking technology measures where and
how long subjects look at a stimulus, such as
a face or a video of an interaction
– Brain imaging technology shows brain activity
in response to a particular stimulus or
situation
Copyright © 2017 Cengage Learning. All Rights Reserved.
Testing Ideas

Research Designs

Copyright © 2017 Cengage Learning. All Rights Reserved.


Research Designs

• There are three main types of research


designs social psychologists use to test
their ideas
– Descriptive
– Correlational
– Experimental

• Let’s start by going over the first type,


descriptive research, on the next slide.

Copyright © 2017 Cengage Learning. All Rights Reserved.


Descriptive Research: 

Discovering Trends and Tendencies

• Goal is to describe people and their thoughts, feelings, and behaviors

• Descriptive research can be accomplished by:


– Observational studies (simply observing people)
• Ex: researchers watch and take notes on children’s interactions on the playground to describe
the frequency and severity of bullying.
– Archival studies (studying records of past events and behaviors)
• Ex: examine existing records of past events, such as newspaper articles, medical records,
crime statistics, hits on a website
• An advantage is that since the information is secondhand, the researcher can be sure that
they didn’t influence participants’ behavior by their presence
– Surveys (like opinion polls which can be conducted in person, by phone,
online)
• But recall that the wording and question order can influence self-reports!

• This method can test questions such as:


– Do most people support capital punishment?
– What % of people who encounter a person lying on the sidewalk would help ?
– What do men and women say are most likely to make them jealous?
Copyright © 2017 Cengage Learning. All Rights Reserved.
Observing Basketball: The Ball Don’t
Lie (Hayne & Gilovich, 2010)
Another example of a descriptive research question:
are professional basketball players more likely to miss
a free throw if they were awarded the shot because of
an obviously bad foul call by the referee?

Method: Researchers observed 102 games. Noted when


obviously bad calls were made and observed whether
the subsequent free-throw shooter made/missed the
shot.

The graph compares free-throw success rates for:


1.The average NBA player (first bar)
2. Season average for the players in the study (2nd bar)
3.The first shots after a bad call (third bar)
4.The second shots after a bad call (last bar)

The results demonstrated the players shot much worse


immediately after benefitting from a bad call.

But if you want more information (like how or why),


you’ll need to conduct a correlational or experimental
study!
Correlational Research: 

Looking for Associations

• Measures the relationship between variables


– How similar or distinct are two different variables?
– How well does one variable predict another variable?
– Ex: is there a relationship between gender and willingness to ask
for help from others? Between self-esteem and popularity?
– Note: in correlational research, researchers aren’t manipulating
the variables, just studying them.

• Correlation coefficient
– Measures the strength and direction of the relationship between
variables
– Ranges from +1.0 to –1.0
– Absolute value indicates how strongly the two variables are related
– The positive or negative sign indicates direction of the relationship

Copyright © 2017 Cengage Learning. All Rights Reserved.


Correlations: Positive, Negative, and
None

A positive correlation indicates that that as one variable increases, so does the
other, and as one decreases, so does the other - the variables go in the same
directions.
-example: as the temperature increases, so does the # of people who buy cold drinks. As the
temperature decreases, so does the # of people buying drinks.

A negative correlation indicates that variables go in opposite directions.


-example: as the temperature increases, the # of people who wear sweaters decreases. As
the temperature decreases, the # of people who wear sweaters increases.

When two variables are not systematically associated, there is no correlation.


-example: no relationship between daily temperature and the # of people who have hiccups
Example of a Correlational Study

• Researchers examined the relationship between the language used


on Twitter and measures of health (from public records).
– Studied 826 million tweets by people in more than 1,300 counties in the U.S.

• Found that communities in which people used angry language also


tended to have greater rates of heart-disease mortality (Eichstaedt
et al., 2015).

• Test yourself: does this finding represent a positive correlation or a


negative correlation?

Copyright © 2017 Cengage Learning. All Rights Reserved.


Correlation Strength

• Take another example: college entrance exam scores correlate


positively with grades.
– Positive direction indicates higher entrance exams are associated with
higher grades and lower entrance exam scores are associated with
lower grades
• The correlation is not perfect, though, because some people with
high entrance exam scores have poor grades and vice versa.
• So, the correlation is less than +1.0 (the highest possible
correlation), but grater than 0 because there is some association
between the two.
– Few variables are perfectly correlated, so most correlation coefficients
do not approach +1.0 or -1.0
– Most have more moderate values like +.39 or -.57
• Remember, the higher the number the stronger the relationship,
regardless of the positive or negative sign, which only indicates
direction!

Copyright © 2017 Cengage Learning. All Rights Reserved.


Advantages and Disadvantages of
Correlational Research

• Advantages
– Can study the associations of naturally occurring variables that
cannot be manipulated or induced
• Like if you wanted to study age, ethnicity, income
– Can examine phenomena for research purposes that is difficult
or unethical to create
• Like research on love, hate, or physical abuse
– Offers freedom in settings in which the variables are measured
• Can be studied in laboratory or in a real-world setting
• One very serious disadvantage
– Correlation is not causation!
– This means, a correlation cannot tell you whether there is a
cause-and-effect relationship between the variables, only that
they are related to each other.
– Due to the third variable problem (see next slide).
Explaining Correlations: Three Possibilities 

AKA The Third Variable Problem

There is a positive correlation between how much


children play violent video games and how aggressively
they behave, but this relationship can be explained in 3
different ways.
1.Playing violent video games causes aggressive behavior
A! B
2.Children who behave aggressively like to play violent
video games B ! A
3.OR there could be a THIRD VARIABLE that causes both
variables such as: Children who have family troubles,
such as parents who are not very involved in the
children’s development, tend to both play violent video
games and behave aggressively. C ! A and B

Remember this when you encounter reports in the media


that suggest cause-and-effect relationships based purely
on correlational research!

Be skeptical and think critically when you encounter this


issue, as many news sources are guilty of this. You will
become skilled at seeing the flaws in media report.
Copyright © 2017 Cengage Learning. All Rights Reserved. Remember: Correlation is NOT causation.
Experiments:

Looking for Cause and Effect

• But researchers do often want to examine cause-and-effect


relationships.
– How? By conducting an experiment.
• Experiments are therefore the most popular method of testing
ideas in social psychology.
• Experiments have two essential characteristics:
– Researcher controls the procedures and variables (manipulates the variables
of interest while insuring others remain uniform)
• This means all participants are treated in exactly the same manner, except for the specific
differences the experimenter wants to create.
– Participants are randomly assigned to the conditions of the experiment
• Random sampling versus random assignment
– Random sampling is concerned with how individuals are selected to
be in a study
– Random assignment focuses on how participants are assigned to
different conditions
• Randomly! (i.e. coin flip if there are two conditions)
Copyright © 2017 Cengage Learning. All Rights Reserved.
Correlations Versus Experiments

Copyright © 2017 Cengage Learning. All Rights Reserved.


Laboratory and Field Experiments

• Laboratory settings
– Most experiments in social psychology are conducted in a
laboratory setting of a university
– It gives researchers control over the environment
– Measure participants’ behaviors precisely
– Keep conditions identical for participants
• Field research
– Real-world settings outside of the laboratory
– Advantage: people are more likely to act naturally
– Disadvantage: experimenter has less control; cannot ensure
consistent conditions

Copyright © 2017 Cengage Learning. All Rights Reserved.


Types of Variables

• Independent variables
– Variables that the researchers manipulate
• Dependent variables
– Variables that are measured to determine whether manipulation of the
independent variable caused a change
• Subject variables
– Variables that characterize pre-existing differences among study
participants; can’t be manipulated or randomly assigned
– Ex: gender, ethnicity

• Note: if a study includes subject variables, but no true,


randomly assigned independent variable, it is NOT a true
experiment!
– Why? Because random assignment is one essential characteristic of an
experiment.
Copyright © 2017 Cengage Learning. All Rights Reserved.
Experiment: Influence of Others’
Reactions
This graph shows the results of research in which
participants saw different versions of a tape of a 1984
presidential debate between Ronald Reagan and Walter
Mondale.

During the debate, Reagan had delivered a pair of witty


one-liners that elicited a positive audience reaction.

Participants who saw an unedited version of the tape and


participants who saw a version with the jokes and the
audience reaction edited out judged Reagan’s performance
similarly.

Participants who saw a version with the jokes left in but


the audience reaction edited out (suggesting that the
audience didn’t find the jokes funny) rated Reagan much
Independent Variable: Type of tape watched; 3 different conditions
more negatively.
1. Unedited tape
2. Jokes and reaction edited out
3. Jokes left in but reaction edited out
Dependent Variable: Judgment of Reagan’s performance on a scale from 0 (terrible) to 100
(excellent)
Copyright © 2017 Cengage Learning. All Rights Reserved. Adapted from Fein, Goethals, & Kugler,
2007.
Statistical Significance and
Replications

• Statistical analyses allow researchers to determine the likelihood that


the results could have occurred by random chance
– For example, in the presidential debate experiment, the average rating in the
unedited version was 66, and it was 49 in the edited version with reaction removed.
– Is the difference between 66 and 49 large enough to be meaningful, or could this
difference simply be due to chance?

• The standard convention is: if 5 or fewer times in 100 possible


outcomes, results are “statistically significant” and should be taken
seriously.
– But does not provide absolute certainty
– There’s still a 5% chance that the findings occurred by chance

• Growing emphasis on replicating research findings and using


alternative statistical techniques.

• Remember this when encountering reports in the media as well!


Copyright © 2017 Cengage Learning. All Rights Reserved.
Internal Validity: Did the Independent
Variable Cause the Effect?

• Internal Validity: reasonable certainty that independent variables


caused the effects on the dependent variable
– If some factor other than the experimenter’s manipulation caused
the effect, it is called a confound and is a serious threat to
internal validity

• Control groups are used to rule out alternative explanations for results
– Participants who experience all the procedures except the
experimental treatment.

• “Blind” studies minimize experimenter expectancy effects


– Results you find in your experiment may be produced by your own
actions rather than the independent variable
– Best way to protect against these expectancy effects is to keep
experimenters uninformed or “blind” about who is assigned to
which condition
Copyright © 2017 Cengage Learning. All Rights Reserved.
External Validity: 

Do the Results Generalize?

• Researchers are also concerned about external validity:


– The extent to which results obtained under one set of circumstances would also
occur in a different set of circumstances
• External validity considerations
– Representative vs. convenience samples
– Participant diversity
– Mundane realism: how much the research setting resembles the
real-world setting of interest
– Experimental realism: how compelling and real the
experimental situation is to the participants
• The majority of social psychologists who conduct experiments emphasize
experimental realism

Copyright © 2017 Cengage Learning. All Rights Reserved.


Deception in Experiments

• Deception: providing participants with


false information about experimental
procedures
– Can add to realism
– Studies have shown participants are rarely
bothered by use of deception
– However, its use does create ethical concerns
• Confederates: people who act like they
are participants but are really working for
experimenter
Copyright © 2017 Cengage Learning. All Rights Reserved.
Meta-Analysis: 

Combining Results Across Studies

• A set of statistical procedures for


examining relevant research that has
already been conducted and reviewed
• Allows combining of the results of
individual studies to measure the overall
reliability and strength of particular
effects

Copyright © 2017 Cengage Learning. All Rights Reserved.


Culture and Research Methods

• Research findings in one culture do not


necessarily generalize to another
• Context of questions and the assumptions
that respondents make can influence
responses
• Effective translations can be difficult
• Multilingual people may think or act
differently based on language and setting

Copyright © 2017 Cengage Learning. All Rights Reserved.


Lost in Translation

While it can be easy to create literal translations, it can be very challenging


to create translations that have the same meaning to people from various
cultures.

The table above presents examples, from signs displayed around the world,
of what can go wrong when simple sentences are poorly translated.
Copyright © 2017 Cengage Learning. All Rights Reserved.
Ethics and Values in Social Psychology

Copyright © 2017 Cengage Learning. All Rights Reserved.


Ethics and Values in Social Psychology

• Researchers have a moral and legal


responsibility to abide by ethical
principles
• The use of deception is a concern in social
psychology
• Ethical reviews are required for social
psychology studies before they are
implemented

Copyright © 2017 Cengage Learning. All Rights Reserved.


Institutional Review Boards and
Informed Consent

• 1974, U.S. Department of Health, Education, and Welfare


established regulations to protect human participants in research
• Created institutional review boards (IRBs) to ensure participants’
welfare is protected
• Researchers must abide by professional code of ethics
• Must obtain informed consent from subjects
– Individuals are asked whether they wish to participate in the
research project and must be given enough information to make
an informed decision
– They must be made aware that they are free to withdraw from
participation at any time
– Some methods may not require informed consent
• Ex: anonymous questionnaires, naturalistic observations, and certain types
of archival research

Copyright © 2017 Cengage Learning. All Rights Reserved.


Debriefing

• At the end of their participation,


participants are informed of:
– The nature of the research
– All procedures, including an explanation of
exactly what happened and why
– The purpose of the research
– Any deception that was used
• Especially important when deception is
used!

Copyright © 2017 Cengage Learning. All Rights Reserved.


Ethics and Consent Online

• Loss of privacy has opened the window for


corporations, marketers, and researchers to
record actions in ways that raise new ethical
questions
• “To Facebook, we are all lab rats” (Goel,
2014)
• Consent given with agreement to terms of
service
– But who reads those?
– Few people pay attention to the fine print when
they begin to use a new app or service.
Copyright © 2017 Cengage Learning. All Rights Reserved.
Values and Science: Ongoing Debates
and New Controversies

• Ethical principles are based on moral


values
• Do values affect science in areas other
than ethical issues?
– Can science be totally unbiased and
objective?
– How should values affect scientific inquiry?
• Self-protection against intentional and
unintentional bias or dishonesty
– Not limited to the social sciences
Copyright © 2017 Cengage Learning. All Rights Reserved.

You might also like