Social Research

Syllabus Policies

o Keep phone silenced and put away
▪ If you anticipate needing your phone (e.g., you are a caregiver) email me.
o Avoid laptop distractions (no email, messaging, social media, homework, work for
other classes, etc.) and close laptop during group activities
o Do not wear headphones
o Do not pack up early
o Remain visibly awake
o Face group members during small group tasks; move desks to be closer (if possible)

AI misuse

you cannot copy/paste output directly from AI, ask AI to do an assignment for you, or use
real-time AI writing assistants like Grammarly or Quillbot. All words that you turn in must be your own,
typed by you.

MIT study showing that using AI for writing may erode critical thinking abilities  

The experiment consisted of a pretest phase, a learning phase, and a test phase.
For the pretest phase, participants in both conditions were given 3 one-step fraction problems without
assistance. After each pretest problem, the correct solution was shown.
After the pretest, the instructions, interface, and AI assistant setup were identical to those in Experiment
1. Participants in the AI condition were presented with a series of 11 fraction problems (Appendix B.2),
with an AI assistant (GPT-5) available in a sidebar. The AI assistant was then removed, and participants
were asked to solve 3 additional fraction problems. Participants in the control condition were presented
the same 14 problems without AI assistance. T

Sociology

study of society; how people interact, how society is organized and how power and inequality shape our lives

Macro-level

focus on institutions, systems, policies, large scale and/or long term processess

• systememic racism, gender pay gap, social movements

micro-level

focus on individual beliefe and interactions between people

Epistemology

Branch of philosophy that explores “how we know what we know”

• Epistemological assumptions shape the research questions we ask, methods we select, and how we analyze data

Qualitative Research

General goal: To analyze data that enable rich description in words or images

types: Interviews, focus groups, field research, material-based (eg. content analysis of historical doc, social media)

Quanitative Research

General goal: Analyze data that can be represented by and summerized into numbers

Types—> Survey research (most comm), experiments, social network analysis

Cross-sectional survey

Collect data from a single point in time

e.g. do older people and younger americans hold different beliefs

Panel Survey

Collect data at multiple time points from the same subjects at each time point

e.g. do ____ beliefs change over their life course

cross sectional pros and cons

• easier and cheaper

• cant make claims like x causes y because data on x and y come from same timepoint

• cant assess within-person change

Panel survey benifits

better causal leverage, can examine how X and Y relate over time

• well suited to assess within-person change over time

Know the characteristics of a good RQ

A clear focsed question that guides a study and can be addressed with the collection and analysis of data

considerations: Interesting? Gap in what research shows? Can you make an argument thats IMPORTSNT? is it FEASIBLE

RQ examines the relationship between 2 things (variables)

How does X relate to why? ***

Flaws of bad RQs

Dont ask Yes/No RQs—> we want to write descriptive predictions, not a single word

DONT ask cause-and-effect (causal) RQ

You cannot say x causes a change in Y uness you establish an association. establiash a time odering in which x occurs before y, rule out other explinations for the change/effect

AVOID CAUSAL RQS BY: avoiding causas words (increase, decrease) use safe words like relationship, association

Theory

logical framework used to explain and interperet social phenomena, or a proposed explanation for how something in the world wors

Empiricism

Using our human senses to systematically gather and analyze data about the world

Theory needs empiricism

theory without empirical evidence= speculations without proof

scientific method= theory +empiricism

inductive

use data to build theory

Bottom up

Theory- building

LEANS MORE QUALITATIVE

deductive

test theory with data

Top-down

theory testing

QUANTITATIVE

variables

a research ready representation of a concept

Independent Variable (X)

Variable hypothesized to lead to or cause variation or change in another variable

Also called predictior, explanitiory variable

Dependent variable Y

variable hypothesized to vary depending on or under the influence of another variable (IV)

• The DV is dependant on the IV

Also called outcome variable

Simple Directional Hypothesis

X—→ Y

Describes how variation on the IV affects the DV either directly/positivly or inversely/negitivly

EX: Having more religious parents will be associated with higher personal religiosity

Mediation hypothesis

Predicts that the IV affects the DV because IV affects an intermediate “mediating variable” that affects the DV. MV is the pathway or mechanis, linking IV to DV

EX: The association between parents religiosity and personal religiosity will be partially explained by exposure to religion in childhood

Moderation/interactions hypothesis

Predicts that the IV’s expected effect on the DV will vary based on another IV; these two IV’s will “interacts” to affect the DV

EX: The association between parents religiosity will vary based on part-child relationship quality

causal hypothesis

prediction that the relationship between two concepts is the result of cause and effect

• RISKY unless ur research design is an experiment

Three central aims of a lit review

1. justify the importance of the study

2. conceptulize/ define key terms

3. stategically synthsize the literature thats most relevant to the RQ

empitical claim

statement about observable world that can be proven/tasted using data—→ Cite empirical claims

Organizing a lit review

organizing by theory

organizing by variables

organizing by theory

helpful for qualitative papers

1. section on broad theory that cann apply to many topics

2. section on empirical findings more closely related to RQ

organizing by variables

for quanitiative papers

x—→y

1. section your DV

2. Section showing how x relates to y in prev research

3. explain how/why x relates to y in the way you predict

4. end with hypothesis

common mistakes in lit reviews

1. Summarizing articles one-by-one
2. Too much detail about each study
3. Overreliance on direct quotes
4. No clear organizing logic or flow
5. Vague fluff or unclear writing
6. Conversational tone
7. Failing to cite things that need to be cited
8. Not using scholarly sources

Conceptualization

specifying precisely what we mean by a term

• moving from doncepts to variables

• defining concepts

note: Conceptualization comes before operationalization because we cant measure something until after we define exactly what were trying to measure

Operationalization

specifying a plan/set of procedures for measuring variables

Variables must vary – you should be able to identify examples of studies that wouldn’t be feasible because there wouldn’t be sufficient variation on the independent variable (E.g., examining how gender shapes mental health using a survey of only women…) 

Variables must vary enough to analyze

• cant examine how gender shapes mental health using a survey of only women because it doesnt vary

sampling frame

list of all the members of your target population, from which a probability sample is drawn

• if your population is people, sampling frame= names and contact info

EX: Population: All college students at Uark

SF: The registars list of currently enrolled students

Probability sampling

type of sampling approach in which every unit in the population has a known chance of being selected, uses random selection, and each unit/ inividuals probability of being selected can be calculated

Unbiased; best suited for statistical or "empirical" generalization? 

4 types

Simple random sample

systematic sample

clister sample

stratified sample

Which type of sampling is most akin to drawing names out of a hat? 

simple random sample

cluster sampling

probability sampling strategy in which researchers divide up the target population into group or clusters. First randomly selecting clusters and then randomly selecting individuals/members within those clusters

Stratefied sampling

Probability sampling strategy in which population is divided into groups and sample members are randomly selected in strategic proportions from each group

Random error (noise)

errors scatter unpredictably, canceling each other out in large samples

Systematic error (bias)

errors consistently pull results one way

• this is the worse type, it cannot be reduced by sample size

coverage error

Bias/error that occurs when the sampling frame doesnt correctly capture the target population

e.g. sampling parents via schools directory—> some parents may nnot be listed, others with multiple kids may be listed twice

sampling error

random error that inevitably occurs when trying to learn about populations by studying samples

increase sample size reduces sampling error

Non-response bias

bias/error that occurs when the mwembers of the population who do take the survey differ from the members of the population who do not take the survey in systematic ways

Social desirability

participants might report positively valued behavior/attitudes rather than truthful responses

• technically a respondent-driven source of error, but amplified by mode/ interviewer effects