SOCI2030: Quantitative Research Methods
Lecture 2: Research Questions
Lai Wei
Feb 11, 2025
*I thank Satoshi Araki for sharing slides
Key Ideas from Last Lecture
• Types of Research
• Empirical:
• Quantitative
• Qualitative
• Mixed
• Non-Empirical:
• Literature Review
• Theoretical
• Formal/Game Theory
• Simulation-Based
• Quantitative
• Survey
• Experiment
• Big-Data
Reviewing Last Week
• Read page 10-16 of the shared research article (Wei, Yao and Zhang
2023)
• Discuss:
• What types of quantitative research methods are used?
• Why do the authors decide to combine these two methods? How do the two
methods complement each other?
survey - small in quantity + social media data - large in quantity [complement each other]
survey is structured, easy to categories ppl’s answers as the questions in survey are
controlled and questions are asked in a way that ppl will understand it + social media data is
varied, different ways for ppl to answer the questions, difficult to control [complement each
other]
The Workflow of Quantitative Research
Overview of Today
• Types of Research Questions
• Descriptive
• Explanatory
• Predictive
• What is a good research question?
• Clarity
• Practicality
• Significance
Types of Research Questions
• Description
• describe or summarize patterns, trends, or characteristics of a population or
phenomenon.
• Usually involves one variable
• Explanation
• Seek to explain relationships between variables (often cause-and-effect).
• Involves two (sometimes more) variables
• The MAINSTREAM in the current paradigm
• Prediction
• Aim to forecast future occurrences or outcomes based on known indicators.
• Involves heavy modelling, sometimes machine learning
Types of Research Questions
Types of Research Questions
• Descriptive
• Sociology Examples:
• “What is the average income of college graduates in HK?”
• “How many hours per week do working adults spend on social media?”
• Data & Analysis:
• Quantitative measures like counts, percentages, or averages.
• No attempt to determine cause-and-effect relationships.
Types of Research Questions
• Explanatory
• Sociology Examples:
• “Does increasing police presence reduce neighborhood crime rates?”
• “How do family income and parental education jointly affect a student’s
academic achievement?”
• Data & Analysis:
• Inferential statistics (e.g., regression, t-tests, ANOVA).
• Focus on whether one variable causes changes in another variable.
Types of Research Questions
• Predictive
• Sociology Examples:
• “Which factors best predict whether a teenager will attend college?”
• “How can we predict future population shifts in metropolitan areas?”
• Data & Analysis:
• Emphasize model-building and validating predictions.
• Often involves machine learning models or advanced regression techniques.
• In general, less common than the other two.
Types of Research Questions
• One research may involve multiple research questions, with one of them
being the focus
• Suppose you want to study social media and mental health among college
students
• Descriptive:
• How many hours do college students spend on social media per day?
• Explanatory:
• Is increased social media usage associated with anxiety or depressive symptoms?
• Predictive:
• Can we predict the high-risk students in terms of mental issues based on their social
media habits?
Example
• Read the home ownership paper briefly
• Discuss:
• What is the main research question of the paper?
• Which part of the result (figure/table) answers descriptive questions?
• Which part of the result (figure/table) answers explanatory questions?
• Describe their main finding (answer to their research question)
main research question: how home ownership affects life satisfaction and subjective class identification in Hong
Kong (two significant predictors of individual well-being and social stability) [explnatory questions]
page 12 answers the describtive questions
with the following figures answering the question trying to find a relationship between the cause and effect
Main finding: strongly suggest that homeownership has positive causal effects on life satisfaction and subjective class
identification
On Explanatory Research
• Most often, explanatory research concerns two variables, an
independent variable and a dependent variable
• Also known as predictor & outcome
• Or treatment & outcome (in a strict causal setting)
• Independent variable causes the dependent variable
X Y
On Explanatory Research
• The case that involves a third variable
• Z can be a confounder, that is a common cause of X and Y
• We have to control for Z if we want to understand the relation
it is important to recognise the confunder, cause if we don’t control for the
between X and Y confounder, we’ll misinterpret the relationship between X and Y
Confunder is the cause of X and Y
X Y
if we observe that college graduates have higher income than non-college graduates, it doesn’t mean its a cause and effect relationship
it can be due to richer family backgrounds (confounder), even if those children didn’t go to college, they would still earn more since their
parents are richer
mobility is another (coufounder)
• Examples: college degree and income, what are confounders
On Explanatory Research
• The case that involves a third variable
• Z can be a mediator, that is the mechanism between X and Y
• The relation between X and Y happen through Z
could be due to human capital, learn all kinds of things in college, which would lead
to higher income due to more skills
could also be due to social network, getting to know more people and making
friends, which can lead to higher income since they might find better jobs
X Z Y
• Examples: college degree and income, what are mechanisms
On Explanatory Research
• The case that involves a third variable
• Z can be a moderator, that is the relation between X and Y is different
by Z going to HKU is different for local residents and international
students and its different for people from different backgrounds
X Y
• Examples: college degree and income, what are possible moderators
a variable can both be a confounder and moderator, and be a moderator and mediator, but cannot be confounder and mediator
Exercise
• Match the following 5 variables:
• Social Media Use Independent Variable
• Academic Performance Dependent Variable
• Stress Level Confounder
• Sleep Quality Mediator
• Family Background Moderator
Use these 5 variables:
What is your research question (independent and dependent variables)?
What is a possible confounder?
What is a possible mediator?
What is a possible moderator?
Note on Description
• Majority of social science research questions are explanatory
• Description, or uni-variate analysis, is usually a by product.
• But good descriptive research can be striking too.
Note on Description
Note on Prediction
Salganik et al. 2020; PNAS
What is a good research question?
• There are many ways to summarize what a good research question entails.
• I focus on the following three concepts
• Clarity eg. how does tech influence society
- vague
• Uses precise language; avoids vague terms - need to specify what type of tech and what society
• Narrow enough to be answerable within time and resource constraints
• Practicality
• Can be addressed using available methods and data
• Significance
• Social/real-life significance
• Academic significance: addresses gaps in the literature
What is a good research question?
• Clarity:
• Clear, unambiguous, focused
• In explanatory research questions, specify your independent variable and dependent
variable
• These variables should be reasonably measurable
• Vague example:
• How does technology change the society?
• Exercise: What are better alternatives?
• What is wrong with vague/broad questions?
• We are tied by our limited focus, resources, time.
• Unmeasurable example:
• How does moral virtue affect political decision making?
What is a good research question?
• Practicality
• Whether the research question can realistically be answered given
available resources, time, methods, and data.
• A practical question takes into account feasibility: data availability,
access to participants, ethical considerations, and the methods
needed.
• Discuss the practicality of the following research question:
• "How does daily social media use affect academic performance among high
school students in an urban school district?"
What is a good research question?
• Feasibility Check List
• Data Access:
• Can we access or collect the required data?
• Measurement:
• Are the key variables clearly defined and measurable?
• Tools/Methods:
• Do we have the tools and expertise to analyze the data?
• Scope and Timeline:
• Is the scope narrow enough to be completed within the available time and resources?
• Ethical and Logistical Feasibility:
• Are there any ethical or logistical challenges?
• Resource Availability:
• Do we have the financial, human, and technical resources required for this research?
What is a good research question?
• Significance:
• Social significance: this is something important for the society
• Academic significance: addresses a research gap
• Answering new questions, but stay in dialogue with old researches
• Social significance:
• Think of who might be affected by your research
• The public, policy-makers, educators, communities
• What is the social significance of the following example:
• What is the effect of sleep quality on academic performance among college
students?
What is a good research question?
• Academic significance:
• Filling a gap: where there is limited or no research, conflicting findings, or
underexplored topics within a field.
• No research on an important thing
• There is research but with conflicting findings
• There is research with robust findings, but missed some important aspects
• More on this next week
• Try to identify gaps
• “Studies have documented the negative impact of income inequality on social
trust in Western societies”
What is a good research question?
• Try to identify gaps
• “Studies have documented the negative impact of income inequality on social
trust in Western societies.”
• What are possible research questions that extend the earlier finding?
• Different context?
• Change the independent variable?
• Change the dependent variable?
• Consider mediator?
• Consider moderator?
Practice with your own research question
• What is your research topic?
• Or social issues you are interested in
• Why is it important?
• Academically/practically/socially
• What is known and what is not?
• Or any gap/puzzle
• What is your research question?
• Consider the explanatory format, “How does X affect Y?”
Write up your ideas out and discuss with peers!
Practice with your own research question
• Discuss with your peer and convince them that your research is
significant.
• Refine your research questions together.
• To be shared later.
Overview of Today
• Types of Research Questions
• Descriptive
• Explanatory
• Predictive
• What is a good research question?
• Clarity
• Practicality
• Significance
• Next week:
• Literature review (more on finding the gap)
Check-Out Form