How to Write a Research Question: Types, Examples, and Tips
A research question is the single sentence that defines what your study is trying to find out. It shapes every decision that follows — your methodology, your literature search, your analysis, and your conclusions. Writing a strong research question takes more effort than it looks, but a clear, focused question is the single best investment you can make before beginning any research project.
What is a research question?
A research question is a focused, specific inquiry that your study is designed to answer. It sits at the heart of any academic paper, thesis, or dissertation — distinct from a topic (which is broad) and distinct from a hypothesis (which is a proposed answer). The research question defines the scope of your inquiry and determines what counts as relevant evidence.
Good research questions arise from gaps in the existing literature, unresolved debates, real-world problems that need evidence, or the desire to test whether a well-established finding holds in a new context. Before you can write one, you need enough background reading to know what is already known and where the knowledge ends.
Characteristics of a good research question
The most frequently cited framework for evaluating research questions is the FINER criteria — Feasible, Interesting, Novel, Ethical, and Relevant. Beyond FINER, four qualities should be present in every well-formed research question:
Clear
Every word in the question should have an unambiguous meaning. Avoid vague terms like "impact," "affect," or "improve" unless you define them. A reader should be able to read the question once and understand exactly what the study will investigate.
Focused
A good research question addresses one relationship, one population, or one phenomenon. Questions that try to cover too many variables at once cannot be answered by a single study. If you find yourself using "and" to join two separate inquiries, split them into two questions.
Complex
A strong research question cannot be answered with a simple "yes" or "no," nor can it be answered by looking up a single fact. It requires data collection, analysis, and interpretation. Questions that can be answered by a quick Google search do not belong in academic research.
Arguable
The question should be genuinely open — there must be more than one reasonable answer. Questions about settled facts are not research questions. The best questions exist in contested territory where evidence is incomplete, contradictory, or missing entirely.
Types of research questions
Research questions fall into three broad families. The type you choose determines your methodology — so it is worth being deliberate about which category your question belongs to.
Descriptive questions
Descriptive questions ask what something is like. They seek to describe a phenomenon, population, or situation without establishing causal relationships. These questions suit surveys, observational studies, and qualitative methods.
What are the most common study strategies used by first-year university students in the UK?
How do small-business owners in rural communities describe their experience of accessing credit?
What proportion of adults aged 18–35 in the United States report sleeping fewer than seven hours per night?
Comparative questions
Comparative questions ask how two or more groups, conditions, or time periods differ on a measured variable. They imply that there are at least two distinct groups to compare but do not assert that one caused a difference in the other.
How do reading comprehension scores differ between students taught with phonics-based instruction and those taught with whole-language methods?
Do female and male managers differ in the leadership styles they adopt in high-pressure environments?
How does the prevalence of type 2 diabetes compare between urban and rural populations in sub-Saharan Africa?
Causal (explanatory) questions
Causal questions ask why something happens or what the effect of a particular intervention or variable is. They are the most ambitious type and typically require experimental or quasi-experimental designs to answer rigorously. Causal language — "cause," "lead to," "effect of," "impact of" — signals this type of question.
Does mindfulness-based stress reduction (MBSR) reduce self-reported anxiety scores in adults with generalized anxiety disorder?
To what extent does early childhood exposure to bilingual environments affect executive function in five-year-olds?
What is the effect of a minimum-wage increase on employment rates in the food-service sector?
The PICO framework
The PICO framework is a structured tool widely used in health and clinical research to ensure that a research question contains all the components needed to design a study and conduct a systematic literature search. PICO stands for:
| Letter | Element | What it defines |
|---|---|---|
| P | Population / Patient / Problem | Who or what is being studied (age group, diagnosis, setting) |
| I | Intervention / Exposure | The treatment, exposure, or factor under examination |
| C | Comparison / Control | What the intervention is being compared against (placebo, standard care, another group) |
| O | Outcome | The measurable result you are interested in |
Some versions of the framework add a fifth element — T for Time — making it PICOT, which is useful when the duration of follow-up is important to the study design.
P: Adults aged 40–65 with a diagnosis of hypertension
I: Daily aerobic exercise (30 minutes, five days per week)
C: Standard care with no structured exercise program
O: Systolic blood pressure at 12 weeks
Research question: In adults aged 40–65 diagnosed with hypertension, does a 12-week structured aerobic exercise program reduce systolic blood pressure compared with standard care alone?
PICO is most at home in clinical and health sciences but the same logic — specify the population, what varies, what it is compared against, and what you measure — applies to social-science and educational research too.
Examples across disciplines
Good research questions look different depending on the discipline. The following examples show how the same underlying principles — focus, clarity, complexity, answerability — play out across fields.
Psychology
How does attachment style in early childhood predict romantic relationship satisfaction in early adulthood?
Education
To what extent does formative assessment feedback improve academic self-regulation in secondary-school students?
Sociology
How do first-generation college students navigate institutional belonging in selective universities, and what social factors mediate this process?
Environmental science
What is the effect of urban green-space coverage on ambient temperature and resident-reported heat stress in cities with populations above one million?
Economics
Does access to mobile-payment platforms increase household savings rates among low-income populations in East Africa?
History
How did the introduction of the printing press change the circulation and authority of medical knowledge in sixteenth-century Europe?
Narrowing a broad topic
Most research questions begin as topics — broad areas of interest that need to be narrowed before they can guide a study. The process of narrowing involves adding specificity along several dimensions: population, time period, geography, variable, and context.
A useful technique is to move through a series of progressively sharper questions, each more focused than the last:
Topic (too broad): Social media and mental health
First narrowing: Does social media use affect the mental health of young people?
Second narrowing: Does daily Instagram use affect self-esteem in adolescent girls?
Final research question: Does daily passive Instagram use (scrolling without posting) predict lower self-esteem scores in girls aged 13–17, compared with active use (posting and commenting)?
At each stage, the question gains population specificity (adolescent girls aged 13–17), variable specificity (passive vs. active use), and a measurable outcome (self-esteem scores). By the final version, a researcher knows exactly what to measure, who to recruit, and how to analyse the data.
Strategies for narrowing
The following strategies help move from a broad area to a workable research question:
- Conduct a preliminary literature review to identify what is already known and where the gaps are.
- Identify a specific population — age group, geographic region, professional group, or clinical diagnosis.
- Specify the time frame — cross-sectional (one point in time) or longitudinal (change over time).
- Define key variables precisely — replace "social media use" with "daily minutes spent on Instagram" or "number of posts published per week."
- Consult your supervisor or a subject librarian — a fresh pair of eyes often spots the assumption that is keeping a question too broad.
Common mistakes to avoid
Asking a yes/no question
A question that can be answered with a binary response does not require research — it requires a lookup. Replace "Does social media affect teenagers?" with "How does daily social media use affect anxiety levels in teenagers aged 14–18?"
Asking too many questions at once
A single study cannot answer five questions. If your research question contains multiple "and"s connecting separate inquiries, pick the most important one and save the others for future work.
Selecting a question that is not answerable with available data
If there is no dataset, no population to recruit, and no feasible way to measure the outcome you have in mind, the question is not yet workable. Feasibility is part of a good research question.
Framing a question that assumes its answer
"Why does mindfulness improve student grades?" assumes the effect exists before investigating it. The correct framing is: "Does mindfulness practice improve academic performance in university students, and if so, through what mechanisms?"
Quick checklist
| Criterion | Question to ask yourself |
|---|---|
| Clear | Would someone outside my field understand exactly what this question is asking? |
| Focused | Does the question address a single relationship, population, or phenomenon? |
| Complex | Does it require data collection and analysis — not just a web search — to answer? |
| Arguable | Is there genuine uncertainty about the answer? |
| Feasible | Can this question realistically be answered given your time, resources, and access to data? |
| Novel | Does answering it add something new to the existing literature? |
| Relevant | Will the answer matter to scholars, practitioners, or policy makers in the field? |
Once you have a research question, you need sources to support it. CiteGenie's Find Source tool searches Crossref, Semantic Scholar, arXiv, and more to find the academic papers that best support your claims.
Find Supporting Sources