MODULE 2 RESEARCH METHODHOLOGY OF SOC. PSYC.

COURSE OBJECTIVES:

 CO 1: Identify definitions of social psychology theories and processes.

 CO 2: Describe how social psychology experiments are conducted to support the theories.
 CO 3: Apply social Psychology theories to social situations in everyday life.
 CO 4: Apply APA-style to write social psychology papers.

MODULE OBJECTIVES:

In this module, you will:

 MO2.1: Explain why social psychologists rely on empirical methods to study social behavior. (CO
2)
 MO2.2: Describe various research methodologies that social psychologists use and their
functions. (CO 2)
 MO2.3: Identify the pros and cons of each research methodology. (CO 2)

READING:

Principles of Social Psychology (PSP), Chapter 1.3 Conducting Research in Social Psychology

LESSON: KEY LEARNING POINTS (ALSO SERVE AS THE STUDY GUIDE)

THE SCIENTIFIC RESEARCH PROCESS (MO 2.1)

In Module 1, we introduced the scientific process. We also called it empirical methods. The scientific
process often starts from making a hypothesis (which is also an educated guess). Then experiments are
conducted to test the hypotheses. The results obtained from experiments then inform the researchers
how behaviors can be predicted or explained, and how to find solutions to problems. This is a recurring
process in which the results then loop back to modify the hypothesis if necessary. With an updated
hypothesis (from previous research), researchers then continue to employ the scientific process to
conduct experiments.

Figure 1. The scientific process employed by social psychologists

You may say, is this necessary? Do you think that social psychology is just common sense? For example,
some people think that they do not need scientific research to tell them that we are attracted to similar
others (i.e., “birds of the same feather flock together”). But how about “opposite attract”? Without a
careful scientific research process, how can we know if humans are attracted to similar others, or
opposite others. You may say, “it depends”. Exactly! We need to conduct scientific research to find out
what situations people may be attracted to the opposite and what situations people may be attracted to
the similar!

People who think social psychology is only a common sense commits hindsight bias. Hindsight bias is
“the tendency to think that we could have predicted something that we probably would not have been
able to predict” (PSP, Chapter 1.3). It is a human tendency to exaggerate, after knowing that something
occurred, how much they could have predicted it before it occurred. (e.g., “I could have told you that”,
Monday morning quarterback, knew-it-all along effect).
WHAT ARE THE RESEARCH METHODOLOGIES THAT SOCIAL PSYCHOLOGISTS USE? (MO 2.2,
2.3)

Research methodologies are like the tools that researchers use. You use different tools when working
on different tasks. Similarly, when researchers attempt to solve different research questions, they use
different methodologies.
Before choosing the right methodologies, the first task is to identify a research question. In a research
question, there are variables, factors or components that you are interested in studying. For example, if
you want to understand suppression on minorities, then you can look into variables such as prejudice,
discrimination, wage gap across racial groups, or intergroup conflicts.
Then, a conceptual definition and an operational definition needs to be developed. A conceptual
definition is the characteristics or contents of your variables. An operational definition is how you are
going to measure the variable. Here is an example,
Research question: What makes people prejudice against racial/ethnic minorities?
Variable: Racial prejudice
Conceptual definition: Unfavorable feeling, thoughts, or behaviors towards a person because of his/her
race/ethnicity.
Operational definition: Agreement to negative descriptions of a racial/ethnic minority member.
Next, we need to choose the right tool (method) to find out the answers as a social psychologist!

OBSERVATIONAL STUDY
The goal of observation study is the description of phenomena. It answers questions such as “how does
a person describe…” “how does a person perceive…” Observation can be naturalistic or ethnographic.
Naturalistic observation is a study that observe behaviors as an outsider of the group. Ethnographic
observation is a study in which the researcher becomes a member of the target group and observes
from the inside.

Here is an example of a naturalistic observational study: Laughing Tram Man has a contagious laugh. In
the video, you can observe how people react to a stranger laughing on the street. To measure the
behaviors, you can count how many times people around the man laugh, or how long people laugh.

However, how can we maintain objectivity when observing the behaviors? Researchers often use more
than one rater (observer). After gathering all observers’ rating, researchers can then calculate inter-
rater reliability (i.e. percentage of agreement among raters).

An example of ethnographic observation is the observation of homeless people by becoming a homeless

person in the community. This short film gives you an example of how to do ethnography.

Other than human behaviors, researchers also observe text or recordings, such as past documents,
magazines, news articles, and social media posts. This is called archival analysis or content analysis. See
what my social psychology students and I published on the Journal of Community Psychology in a study
observing MeToo social movement on Twitter.

What are the pros and cons of observational studies?

Pros: Study phenomena that can’t be studied in a lab (e.g. social movement, work environment).

Cons: Less control over extraneous variables (i.e. error). For example, in the laughing study, you cannot
control if someone comes disrupt everyone’s laugh. Observational studies also cannot demonstrate
cause and effect – it only describes a phenomenon.

CORRELATIONAL STUDY
The goal of correlational study is to predict behaviors. Correlation is the inter-relation between two
variables. Research questions that it addresses are: “What are the relationship between instruction
method (online vs. face-to-face) and students’ success”, “what are the relationship between childhood
trauma and romantic relationship style”, etc.

Correlational studies are often done by using surveys. You probably have done some in SONA
Psychological Research Participation system. In a survey, researchers ask participants various questions
relating to their research questions. Then, using statistical methods, researchers analyze the
relationships between the studied variables. Correlation analysis looks like the pictures below. The first
picture (Figure 2) shows a positive correlation between height and weight (i.e. the taller you are, the
heavier you are). The second picture shows a negative correlation between fitness and percentage of
fat. The more exercise one does, the lower the body fat. Finally, the last picture shows a zero-
correlation, showing that there is no relationship between age and height for people who are over 25
years old.

Figure 2 Positive correlations: variables move in the same direction

Figure 3 Negative correlations: variables move in the opposite direction.

Figure 4 No correlations. The line appears to be flat, showing no relationship between x and y-axis.
How do we interpret correlation coefficient (r)? Specifically, what do a positive r and negative r mean?
When r is close to zero, what does that mean? To understand correlation coefficient, first, the sign
indicates positive or negative direction. This means that if you see a correlation coefficient that is
positive (+r), the relationship between the two variables are positive (Figure 2; when one variable
increases, the other increases). If you see a correlation coefficient that is negative (-r), the relationship
between the two variables are negative (Figure 3, when one variable decreases, the other increases).
When you see a correlation coefficient that is close to zero, that means there is no relationship between
the two variables.

What are the pros and cons of correlational studies?

Pros: Correlation is an efficient way to look at relationship between two variables. It is cost-effective. It
involves little time and money. It also helps to predict phenomenon.

Cons: However, there are some major problems with correlation: It only predicts but cannot draw
causation because of two issues:

1. Directionality: We only know that the two variables are related, but we do not know which
causes which. For example, if you find a correlation between eating chocolate and being happy,
you do not know if chocolate makes people happy, or if mood (being happy) increases the
amount of chocolate one eats. (Also explained in Figure 5)

Figure 5. An example of directionality problem.

2. Third variable problem: Even if you find statistical relationship between two variables, in reality,
there may not be a real relationship between them. For example, it is possible that in reality,
chocolate and happiness are not related. It is money (a third variable) that causes the existence
of both events. Having money makes you happy, and having money allows you to buy more
chocolates. (Also explained in Figure 6)
Figure 6. An example of third variable problem.

In short, in correlation, we know there is a relationship, but we do not know what the relationship
means.

EXPERIMENTAL STUDY
The goal of experimental studies is to find causal relationship/ to answer the question “why”, such as
“Why do people help a stranger?”, “what causes people to like one another?”, etc.

In psychology, the word “experiment” is strictly used to refer to studies that Randomly assign
participants to different conditions and ensure that these conditions are identical except for the
independent variable (we will explain these terms below). We do not use the word “experiment” to
refer to other types of studies (e.g., observation or correlation). Experiments play an important role in
social psychology! The Milgram’s experiment introduced in Module 1 is one of many experiments that
help to build social psychological theories.

What must we do when we conduct an experiment? What are the different components of
experiments? Let’s use this example to explain these experimental terminologies. Let’s say my research
questions is “Does a colorful powerful lecture increase students’ grade”

Figure 7. An example of experimental study.

Conditions: In an experiment, researchers design at least an experimental condition and a control
condition. In my study on PowerPoint Lecture, my experimental condition will be a colorful PowerPoint
lecture, and my control condition will be a black and white PowerPoint lecture.

Independent variable: The independent variable is the variable that researchers manipulate, or the
experimental condition. Therefore, in this case, my independent variable is the Color of the PowerPoint
lecture.

Dependent variable: The dependent variable is the variable researchers measure. In this case, it is the
students’ grades.

Extraneous and Confounding variable: Extraneous variable that is a source of error in your experiment.
Confounding variable is even worse, it is a variable that (unintentionally) changes along with, or instead
of, the independent variable, i.e. changes your experimental results. Extraneous variable should be
avoided, but they may not influence your results. Confounding variable is undesirable and will definitely
influence your results. But regardless, both variables should be eliminated.

In my experiment, I have to make sure the two conditions are completely equal except the independent
variable that I manipulate. That is, I have to make sure the wordings of the PowerPoint lecture are the
same across the two conditions. I should only change the color of the words. I should also make sure
the two classes are equal in size, students’ average knowledge in psychology, etc. If not, then I do not
know whether my independent variable has an effect on my dependent variable, or if these extraneous
variable causes the difference in grades.

Randomly assignment is a method to put participants in either the experimental or the control
condition. By randomly assigning participants to conditions, characteristics of participants between
experimental and control groups can be assumed to be equal. Therefore, extraneous and confounding
variables can be minimized

Random sampling: Random sampling applies to all research design – it is to draw your sample (i.e.
recruit participants) randomly from the population to ensure that the sample being studied is
representative of the larger, broader population under question. In my experiment, I may want to not
only draw social psychology students, but also students from different majors/universities randomly to
make sure the students in my classroom is representative of the students across majors/universities.

Internal validity. Internal validity is achieved when nothing besides the independent variable affect the
dependent variable. In other words, once the experiment is free of error, it is said to have internal
validity.

External validity. External validity is how generalizable the experiment is to the real world. It is achieved
when the results of a study can be generalized to other situations and other groups of people. I can
increase my external validity by conducting my experiment in a real classroom. Also, I should not only
conduct in one class or one university. I should try to conduct the experiment as many subjects and as
many universities as possible.
What are the pros and cons of experimental studies?

Pros: When confounding is well-controlled, it allows conclusions about cause & effect relationships
between variables

Cons: Difficult to control all confounding. It is artificial: More control = father away from reality (i.e.
lower external validity). Also, there are some questions that cannot be tested in an experiment. For
example, child abuse, divorce, death, etc.

META-ANALYSIS
Meta-analysis is a statistical procedure for combining data from multiple studies. Meta-analysis can be
used to identify common effects among different studies. It is best for topics that have already been
widely studied in the field. It is also best for studies that show different results. In other words, if many
studies have been conducted by other people on the topic yet the findings are conflicting, meta-analysis
is a great method!

RESEARCH ETHICS (MO 2.1)

No matter what methods researchers use, they have to abide to the research ethnical guidelines. Two
associations manage ethics compliance in psychological research: Institutional Review Boards (IRB) and
American Psychological Association (APA).

The golden rule to determine whether a study is ethical is that potential benefits must outweigh
potential harm. Consider the two examples below:

Deception study 1: Milgram’s experiment - Participants were told they were giving electric shocks to the
learners, when the learners (confederate) were screaming and even expressed pain

Deception study 2: Public bystander deception study (a field experiment). In a public space, someone
(confederate) screamed for help. Experimenters observe people’s reactions (to help vs. not to help).

Which one do you think is appropriate? Study 1 or 2, or neither? Deception is defined as:

 misinforming the participants about the true nature of the study.

 misleading the participants about the true nature of the study.
 Debriefing (see definition below) is required for a study that involves deception.

Deception is allowed in social psychology because sometimes, if people know the real purpose of the
study, they may behave in a socially desirable way. Therefore, psychologists use mild deception to
avoid problems such as social desirability and demand characteristics. For example, if people know
that the study is about obeying to authorities to kill – they will not do it. If people know that the study is
about helping others – they probably will do it. So in social psychology, researchers often deceive
participants with a cover story.
In the examples above, deception study 2 is deemed acceptable, because the potential harm is little, but
the potential benefits in understanding helping behaviors is tremendous (Remember the gold rule?
Benefits outweigh harms/risks). However, deception study 1 is inappropriate because the potential
harms/risks is way too high – forcing people to kill someone can cause serious PTSD!!!

Other than the gold rule, here are some important ethical components in a study:

Informed consent: Agreement to participate in an experiment, granted in full awareness of the nature
of the experiment, which has been explained in advance.

Debriefing: Explaining to participants, at the end of an experiment, the true purpose of the study.