How to Write Research Questions
Daniel Ha · Seoul National University PhD StudentEvaluate your questions against the FINER criteria, structure them with the
PICO framework, then convert them into testable hypotheses. Choose among four
types — descriptive, comparative, relational, or causal — and the appropriate
methodology will follow naturally from your choice.
Why do research questions matter?
A research question is the compass of your paper, guiding the direction of your literature review, methodology, and analysis.
A vague research question makes it impossible to define the scope of your literature review, leaving you reading papers endlessly. It also causes your methodology to lose focus, and even after collecting data, drawing meaningful conclusions becomes difficult.
A clear research question, by contrast, immediately narrows the papers you need to read, naturally points toward an appropriate methodology, and lets you explain your paper's contribution in a single sentence. Investing one day in crafting your research question can save months during the research process.
What types of research questions are there?
There are four types — descriptive, comparative, relational, and causal — and identifying which one fits your study makes question formulation much easier.
| Type | Purpose | Example | Characteristics |
|---|---|---|---|
| Descriptive | Captures a phenomenon as it is | "What is the current state of AI tool adoption among graduate students?" | Best suited for exploratory research |
| Comparative | Compares groups or conditions | "How do learning outcomes differ between online and in-person instruction?" | Requires comparison groups and statistical testing |
| Relational | Explores associations between variables | "What is the relationship between social media usage time and academic achievement?" | Establishes correlation, not causation |
| Causal | Identifies cause and effect | "Does a mindfulness meditation program reduce academic anxiety?" | Requires experimental design; highest difficulty |
How do you create research questions?
Follow four steps — derive questions from your topic, apply frameworks, formulate hypotheses, and review — to produce clear, testable research questions.
Step 1: Derive questions from your topic
Ask "what," "how," and "why" about your topic of interest to generate at least 5~10 questions. At this stage, focus on quantity rather than quality.
| Topic | What | How | Why |
|---|---|---|---|
| AI education tools | What is the actual adoption rate? | How do they change the learning process? | Why do students use or avoid them? |
Step 2: Apply frameworks
Validate and structure the questions from Step 1 using the FINER and PICO frameworks. FINER filters out questions that cannot be researched, while PICO clarifies the population, variables, and comparison conditions, turning vague questions into forms that a paper can answer.
FINER — Evaluating question quality
| Criterion | Meaning | Self-check question |
|---|---|---|
| F (Feasible) | Can it be done? | Can I answer this with my available time, resources, and skills? |
| I (Interesting) | Does anyone care? | Would someone other than me want to know the answer? |
| N (Novel) | Is it new? | Has this question not yet been adequately answered? |
| E (Ethical) | Is it ethical? | Can I conduct this research ethically? |
| R (Relevant) | Does it matter? | Does it contribute meaningfully to academia or society? |
Example — Applying FINER to "What is the effect of AI tutoring tools on university students' learning?"
| Criterion | Check | Result |
|---|---|---|
| F | Can recruit students from the same university | ✓ |
| I | Hot topic in educational technology | ✓ |
| N | Limited research in the local university context | ✓ |
| E | Requires privacy safeguards but IRB approval feasible | ✓ |
| R | Directly informs improvements in university pedagogy | ✓ |
If any of the five criteria gets a "no," the question needs revision.
PICO — Structuring the question
PICO originated in medicine but is a useful question-structuring tool across the social sciences as well.
| Element | Meaning | Example |
|---|---|---|
| P (Population) | Study population | Third- and fourth-year students at four-year universities |
| I (Intervention) | Intervention / Independent variable | Use of an AI-based writing feedback tool |
| C (Comparison) | Comparison condition | Students receiving only traditional instructor feedback |
| O (Outcome) | Outcome / Dependent variable | Change in academic writing competency |
Combining PICO yields: "Among third- and fourth-year university students (P), does using an AI-based writing feedback tool (I), compared to receiving only traditional instructor feedback (C), lead to a difference in academic writing competency (O)?"
Step 3: Formulate hypotheses
Convert research questions that have passed the FINER and PICO filters into statistically testable hypotheses. Hypotheses are what make it possible to determine which data to collect and how to analyze it.
| Item | Content |
|---|---|
| Research question | What is the effect of AI tutoring tool usage on self-directed learning competency among university students? |
| Research hypothesis (H1) | Students who use AI tutoring tools will demonstrate significantly higher self-directed learning competency than those who do not |
| Null hypothesis (H0) | There will be no significant difference in self-directed learning competency based on AI tutoring tool usage |
A hypothesis should be specific (clear variables and direction), testable (verifiable with data), and theoretically grounded (answering "Why do I expect this outcome?").
Nubint AI's Hypothesis Generator takes your research question, analyzes related literature, and proposes several testable hypotheses along with theoretical rationale and suggested verification methods.
Step 4: Review and revise
Give your research question and hypotheses a final check to confirm they are clear and appropriately scoped for a single paper. Skipping this step risks having to change direction mid-research.
| Checklist item | Check |
|---|---|
| Is the question expressed clearly in a single sentence? | ☐ |
| Are the independent and dependent variables unambiguous? | ☐ |
| Is the scope answerable within a single paper? | ☐ |
| Is the hypothesis stated in a measurable form? | ☐ |
| Does your advisor agree with this direction? | ☐ |
The Hypothesis Evaluator assesses your hypotheses from multiple angles, including the FINER criteria, and provides specific strengths, weaknesses, and suggestions for improvement.
What is the difference between a research question and a hypothesis?
A research question is an open-ended inquiry exploring "what" or "how," while a hypothesis is a predictive statement claiming a specific outcome will occur.
| Aspect | Research Question | Hypothesis |
|---|---|---|
| Form | Open-ended question ("What is...?") | Predictive statement ("...will be...") |
| Research type | Used in both qualitative and quantitative | Primarily used in quantitative research |
| Role | Sets the research direction | Subject of statistical testing |
| Example | "What effect do AI tools have on learning?" | "Students who use AI tools will score significantly higher on exams" |
In qualitative research, a research question alone is often sufficient without a formal hypothesis. In quantitative research, converting the research question into a hypothesis is essential. Nubint AI's Hypothesis Generator can help you derive appropriate hypotheses from your research questions.
What are common mistakes when writing research questions?
The most common mistakes are embedding an answer in the question, leaving variables vague, or setting the scope too broadly.
| Vague Question | Problem | Improved Question |
|---|---|---|
| "What is the impact of social media on society?" | Too broad | "How does Instagram usage frequency affect body image satisfaction among women in their twenties?" |
| "Is remote work good?" | Value judgment | "What is the effect of remote work on job productivity and satisfaction among IT professionals?" |
| "What are students' study habits?" | No variables | "What is the relationship between time management strategies and first-semester GPA among college freshmen?" |
| "If X improves Y..." | Answer assumed | "What effect does X have on Y?" |
A good question includes a specific population (who), clearly defined variables (what to measure), and a bounded context (under what conditions). Vague variables make methodology design and data collection difficult and conclusions harder to draw.
What are tips for writing research questions by field?
Each discipline has its own conventions for research question format and emphasis, so you should tailor your questions to fit your field's norms.
| Field | Key Point | Example |
|---|---|---|
| Social sciences | Choose a theoretical framework first, then derive questions | "What correlation exists between research stress levels and publication productivity among graduate students?" |
| Engineering / CS | Define "performance improvement over existing methods" with measurable metrics | "How does project completion quality differ between synchronous online and in-person instruction among engineering undergraduates?" |
| Medicine / Health | Always apply PICO and address ethical considerations at the design stage | "Does participation in an 8-week mindfulness meditation program significantly reduce academic anxiety scores among master's students?" |
| Humanities | Descriptive and interpretive questions are common; research aims often replace formal hypotheses | "What are the frequency and purposes of generative AI tool usage among humanities graduate students?" |
Summary
The research question is the point from which every activity in your paper originates and returns. Evaluate quality with FINER, structure with PICO, and convert into hypotheses — this process builds an unshakable foundation for your research.
Once your research question is finalized, move on to the How to Design Research Methodology guide to design a methodology that fits your question, and the How to Write a Research Proposal guide to document your full research plan.