Chapter 2: How Sociologists Do Research

Scientific Versus Unscientific Thinking

• Critical thinking skills very important
• In everyday life, our biases easily influence our observations.
• Biases often lead us to draw incorrect conclusions about what we see.
• Ten common errors are regularly found in unscientific thinking.

Types of Unscientific Thinking

1. Tradition: unscientific way of thinking, using historic ways to deal with a phenomenon,
appealing to traditional ways of doing something

2. Authority: thinking something is true because we read it in an authoritative source/hear

it from an expert, because it is an authority figure it must be right

3. Casual observation: carless observation/observers, highly unreliable, not specific

4. Overgeneralization: taking something very specific and applying it to something much

broader

5. Selective observation: unconsciously ignore evidence that challenges our firmly held
beliefs, only acknowledging one variable, the one that best fits

6. Qualification: made in everyday life: exceptions to the rule, qualifying a finding in

qualitative research becomes an issue ex. Someone studied for an exam and did poorly,
must be because there is something going on in their life

7. Illogical reasoning: because it happened so often, something to do with luck

8. Ego-defense: committed to conclusions they reach in their research because of what

they have invested, "I know this area, this is what I do, I can't be wrong"

9. Premature closure of inquiry: way of thinking that involves deciding that all the relevant
evidence has been gathered on a particular subject, I figured out the answer, we don’t
need to research any further, I found the final answer

10. Mystification: committed to discovering observable causes of observable effects, find no

rational explanation, contributing anything we don’t know to some mystical power, ex:
because we can’t explain the social phenomenon, it must be God
Michael Shermer: Why people believe weird things
• The idea between data and research
• Mystification example

The research cycle:

• What are the theories

• What is the research of things

Main Methods of Sociology

The four main methods in sociological research:
1. Experiments - almost never really used
2. Surveys
3. Field Research
4. Analysis of existing documents and official statistics

Experiments
• Quantitative
• Highly expensive
• Not used frequently, because you are looking at the social world it is extremely difficult to set
up the social world in a specific way, you have to control variables and because it is a
social world and you don’t control every specific thing it might not be accurate
• Experiment: artificial situation that allows us to focus on cause and effect
• Uses randomization: assigning each individual by chance processes

Steps in a Simple Experiment

• Control Group - do not introduce independent variable
• Experimental Group - introduce independent variable
1. Randomize assignment of subjects to group
2. Measure dependent variable
3. Do not introduce independent variable/introduce independent variable
4. Measure dependent variable again

Surveys
• Quantitative
• Ask people questions about their knowledge, attitudes, or behavior either in a face-to-face or
telephone interview or in a paper-and-pencil format.
• Numerous amount of ways, online, telephone, paper and pencil, online surveys, texting
surveys

Closed- and Open-Ended Questions

• Questionnaires may contain two types of questions:
• Close ended: Yes or no, rate from 1-10, how old are you, there is no interpretation
• Open ended: opinion on something, they have a say, every response will be dependent on
how the person feels etc, question like: How do you access the internet? Someone can
answer where, when, why, or how …

Field Research
• Qualitative
• Field research: observing people in their natural environment
• Strategies used in field research:
- Detached observation: watching but do not become a part of it
- Participant observation: becomes a member of that group

Detached Observation
• Observing what was happening in that set
• Watching in natural setting but you do no become a part of it
• Classifying and counting behavior of interest according to predetermined scheme
• Methodological concerns:
1. Reactivity: change how you behave because you see the person is there
2. The meaning of the observed behavior may remain obscure to the
researcher.

Participant Observation
• Involves carefully observing people’s face-to-face interactions and participating in their lives
over a long period of time.
• Researcher becomes a member of that group
• Purpose: join the group and become a part of it
Participant Observation: Strengths and Weaknesses
• Strengths: Allows researchers to develop a deep and sympathetic understanding of the way
people see the world.
• It is especially useful in the “exploratory” stage of research.
• Weaknesses/issues: if people aren't consent to it, it might cause issues and break ethical
views, expensive and time consuming
• Reactivity: people become less open once they find out someone is there, you want to
become a member of the group so people don’t react differently and instead act the
same as they would from day to day

Analysis of Existing Documents and Official Statistics

• Census data, criminal records, police records, court records
• Existing documents and official statistics are created by people other than the researcher for
purposes other than sociological research.
• Strengths: inexpensive, easy to obtain, highly reliable,
• Weaknesses: Not created with the researchers’ needs in mind. Often contain biases that
reflect the interests of the individuals and organizations that created them.

Strengths and Weaknesses of Four Research Methods

Key ideas - KNOW THESE

• Operationalization - question on EVERY EXAM: taking something abstract and making it
concrete - something you can measure; ex: how to operationalize "Intelligence", ask
questions about IQ, if they ever failed a grade
• Variables
• Reliability and validity
• Relationships and causation
• Sampling
• Ethics
Variables
• Dependent variable: always changing, depends on something else, what happens after
(things that can never ever be dependent variables: gender, time, age)
• Independent variable: what causes the change

Reliability and Validity

• Reliability: does your study do what it says it does consistently
• Validity: does what we say what we're measuring actually measure that, am I actually
measuring what I'm saying I'm measuring

Threats to Validity
1. exclusion of part of the population from a sampling frame, excluding a particular group (ex:
only doing online surveys, phone surveys)
2. refusal (African American groups) /inability of some people to participate in the survey
(people may have hearing skills, speech skills)
- High income people and low-income people tend to not reply
3. unwillingness of some people to answer questions truthfully/honestly (context dependent)
4. not giving people to say no (survey design), confusing people, leading people, anything
where nature of the survey influences how people reply

Measurement as Target Practice

Reading Tables
• One of most useful tools for analyzing survey data is the contingency table:
• A cross-classification of cases by at least two variables that allows researcher to
see how, if at all, variables are associated.
• Allows researchers to examine effects of control variables on original association.

Watching TV and Approval of Violence (in percent)

Determining Causes
• To establish whether an independent variable causes change in a dependent variable,
researchers must satisfy three criteria:
1. Relationship test: whether it establishes that the variables are in fact associated. Does
the value of one variable change based on another variable
2. Sequencing: which one happens first, which one affects the other, can we reasonably
argue does one happen before the other
3. Non-spuriousness: Connections between variables and people always exist in a context.

Relationship Test and Sequencing

1. A relationship: exist when two things are systematically connected
2. Relationship test: whether it establishes that the variables are in fact associated. Does
the value of one variable change based on another variable
3. Sequencing: which one happens first, which one affects the other, can we reasonably
argue does one happen before the other
4. Non-spuriousness
- Connections between variables and people always exist in a context.
- Control variables: identify the context for the relationship between
independent and dependent variables
- Is used to control for spurious relationship: relationship appears to exist,
then you introduce a control variable and relationship that seems to exist
disappears
How Spurious Relationships Occur

How Authentic Relationships Occur

Probability Sample
• Researchers have to choose respondents at random, and the chance of choosing an
individual must be known and greater than zero.
• A sample with these characteristics is a probability sample.
• A sampling frame: list of people that are potentially able to participate in it

Sample Size and Statistical Significance

• Large samples give more precise results than small samples do.
• Random sample of 1500 people typically will give acceptably accurate results.
• This level of accuracy is called the margin of error - closer to margin size the smaller your
errors are
The Margin of Error in a Sample

Ethical Considerations
• Treatment of subjects Is needed for respecting rights of research subjects, including:
I. Right to Safety
II. Right to Privacy
III. Right to Confidentiality
IV. Right to Informed Consent - when they know what is going to happen in the research
• Treatment of research results Concern about plagiarism, especially with spread of the
internet.

EXAMPLE OF SOMETHING ON FIRST TEST

• A researcher wants to measure the impact of providing school lunches for children living in
low income areas. She wants to understand the connection between providing meals for
students and their academic performance. She provides lunches to the school for one
year and compares their grades at the end of the year to another school in the area.
• Ask 5 multiple choice questions that you have to answer
- Variables? Independent: the lunches- is the lunches changing the grades?
Dependent: marks student is getting, second school = control group, has to
be similar to the other school
- Reliability and validity?
- Relationships and causation?
- Sampling?
- Ethics?
- CONTROL VARIABLES - to show that relationship that is already happening is
authentic