How to Write a Research Question: Examples and Tips
How to write a research question is the point where a broad interest becomes a specific investigation. A topic such as student sleep, generative AI, nurse burnout or climate communication identifies an area of interest, but it does not yet tell you exactly what evidence you need or what your paper is trying to find out.
A useful question gives the project direction without writing the answer in advance. It helps you decide what to search for, which sources or data matter, what belongs outside the scope, and what kind of analysis will be needed. If the question is vague, the research usually becomes vague too.
This guide begins where the EssayEco cluster on how to choose a research topic ends. It focuses on turning a workable topic into a clear research question. For the wider research process, use academic research and methodology; for the complete writing workflow, see how to write a research paper.
What Is a Research Question?
A research question is the specific question that a research project or research-based paper is designed to answer. It gives the inquiry a center. George Mason University’s research-question guide describes an effective question as clear, focused, concise, complex enough to require analysis, and arguable rather than answerable by a simple fact.
The wording varies by discipline. A history paper may ask how a particular policy was represented in selected sources; a psychology project may ask whether two measured variables are associated; a nursing study may compare outcomes across interventions; and a qualitative project may ask how participants experience a process. The shared requirement is that the question can guide a real investigation.
| Term | What it does | Example |
| Topic | Defines the broad or focused area of inquiry. | Late-night social-media use and sleep among first-year students |
| Research problem | Explains the issue, gap, inconsistency or practical concern that makes inquiry useful. | Students report poor sleep, but the role of late-night social-media habits in this population needs clearer investigation. |
| Research question | States the exact question the project will answer. | How is late-night social-media use associated with self-reported sleep quality among first-year university students? |
| Hypothesis | States a testable expected relationship or difference when the design calls for one. | Greater late-night social-media use will be associated with poorer self-reported sleep quality. |
What Makes a Good Research Question?
A strong question is not simply a sentence that ends with a question mark. It must do several jobs at once: define the inquiry, fit the assignment, leave room for analysis, and remain answerable with the evidence and resources available.
| Quality | What it means | Quick test |
| Clear | Key terms and relationships are understandable. | Could a reader explain what the project will investigate after reading the question once? |
| Focused | The scope is narrow enough for the available length and time. | Does the question avoid trying to study several populations, outcomes or eras at once? |
| Researchable | Evidence, scholarship, texts or data can be used to answer it. | Can you name realistic sources or data that would help answer the question? |
| Analytical | The answer requires interpretation, comparison, explanation or evaluation. | Would a quick factual lookup be insufficient? |
| Feasible | The study can be completed with available access, time, skills and permissions. | Could you actually obtain the evidence the question demands? |
| Neutral enough to investigate | The wording does not assume the conclusion. | Does the question avoid loaded terms such as “prove,” “obviously harmful,” or “failed”? |
| Relevant | The answer has a reason to matter in the assignment or field. | Can you explain what understanding the answer would add? |
| Practical rule
If you already know exactly what the answer must be before you investigate, check whether you have written a claim instead of a research question. The question should guide inquiry, not force evidence toward a predetermined conclusion. |
How to Write a Research Question Step by Step
Step 1: Start With a Workable Topic, Not a Whole Subject
Begin with a topic that has already been narrowed enough to search. “Education,” “mental health,” or “climate change” are subject areas, not useful starting points for a research question. A better starting point might be first-year students’ use of generative-AI feedback, workplace support and burnout among early-career nurses, or responsibility framing in selected climate-news coverage.
If your topic is still too broad, return to the completed guide on how to choose a research topic before trying to perfect the question. Question writing cannot compensate for a topic that still contains several different projects.
Step 2: Do Preliminary Research Before Fixing the Wording
Run a small number of purposeful searches and read enough credible material to learn the field’s language. Look for recurring variables, debates, concepts, populations, methods and gaps. The purpose is not to complete the literature review; it is to discover what kind of question the existing evidence can support.
George Mason recommends preliminary research before finalizing the question because early reading shows what researchers are already discussing and helps expose useful lines of inquiry. A question written without this step may use the wrong terminology, repeat an easily answered issue, or depend on evidence that is not available.
Step 3: Identify the Problem, Tension, Relationship or Unknown
Ask what is actually uncertain or worth examining inside the topic. Are two sources disagreeing? Is there a relationship that needs testing? Do experiences differ between groups? Has a policy produced uneven results? Is a text being interpreted through a particular lens? Is a local population missing from a broader discussion?
| Topic | Possible researchable angle | Question direction |
| Generative-AI feedback | Students may value speed but distrust accuracy. | Experience / perception |
| Nurse burnout | Workload and supervisor support may relate differently to burnout. | Relationship / comparison |
| Climate-news framing | Two outlets may assign responsibility differently. | Comparative / interpretive |
| Online discussion boards | Participation may vary with instructor facilitation. | Association / explanatory |
| Community health program | Outcomes may differ before and after implementation. | Evaluation |
Step 4: Decide What Kind of Answer You Need
The wording becomes easier once you know the intellectual job of the question. Not every discipline uses the same labels, but several broad question types appear repeatedly. These are planning categories, not rigid boxes. A real project may combine more than one.
| Question type | Typical purpose | Useful openings |
| Descriptive | Describe characteristics, patterns, frequency or experiences. | What…? How common…? What patterns…? |
| Comparative | Examine differences or similarities between groups, texts, policies or conditions. | How does X differ from Y…? What differences…? |
| Relational / associational | Examine whether variables or conditions are related. | What is the relationship between…? How is X associated with Y…? |
| Explanatory | Investigate why or how an outcome, process or pattern occurs. | How does…? Why does…? To what extent does…? |
| Exploratory / interpretive | Investigate meanings, experiences, perceptions or under-examined phenomena. | How do participants experience…? How is X represented…? |
| Evaluative | Assess effectiveness, implementation, value or outcomes. | How effective is…? What changes followed…? |
Question type should eventually align with the methodological approach. The research design guide explains qualitative, quantitative and mixed-methods choices in detail. At this stage, you only need enough alignment to avoid asking for evidence your assignment or planned design cannot produce.
Step 5: Add the Boundaries the Question Actually Needs
A question becomes focused by specifying only the boundaries that matter. Depending on the project, those boundaries may include population, setting, location, time period, text, policy, exposure, intervention, comparison, outcome or theoretical lens. Do not add every possible detail merely to make the question sound technical.
| Boundary | When it helps | Example |
| Population | The answer may differ across people or groups. | First-year university students |
| Setting / location | Context is likely to affect the phenomenon. | Public universities in one region |
| Time period | The issue changes historically or the dataset needs a defined window. | News coverage during a selected election period |
| Exposure / intervention | The project examines a specific condition or action. | Late-night social-media use |
| Outcome | The project examines a defined result. | Self-reported sleep quality |
| Text / material | Humanities inquiry needs a bounded evidence set. | Two campaign speeches or one novel |
Step 6: Draft an Open Question
Write several versions rather than trying to produce the final sentence immediately. “How,” “why,” “what,” “which,” and “to what extent” can all work, depending on the study. Avoid assuming that every good question must begin with “how” or “why.” A descriptive study may need “what,” while a comparative study may need a direct comparison.
| Weak first draft | Why it is weak | Improved direction |
| Does social media cause bad sleep? | Loaded wording, causal claim, vague population and outcome. | How is late-night social-media use associated with self-reported sleep quality among first-year university students? |
| Is AI good for students? | Binary, undefined “AI,” undefined students and undefined benefit. | How do first-year composition students perceive the usefulness and limitations of generative-AI feedback during revision? |
| Why are nurses burned out? | Too broad and assumes one general explanation. | What is the relationship between perceived workload and burnout among early-career medical-surgical nurses? |
| How does the media affect climate change? | Unclear mechanism and impossible scope. | How do two national newspapers frame individual versus institutional responsibility in climate-change coverage during a defined period? |
Step 7: Check Whether the Evidence Can Answer It
A well-worded question can still be unusable if the necessary evidence is inaccessible. Ask what an answer would require. If the question asks about attitudes, you may need interviews, surveys or existing datasets. If it asks about historical representation, you need an appropriate corpus of primary sources. If it asks whether an intervention causes an outcome, the design must support a causal conclusion.
This is the point where wording and design begin to constrain each other. Do not use causal words such as “causes,” “leads to” or “impact” merely because they sound strong. If the available study can only identify an association, the research question should not promise causation.
Step 8: Run a FINER Feasibility Check
For projects that involve original research, FINER is a useful second-stage test. The University of Melbourne’s research-question guidance summarizes FINER as feasible, interesting, novel, ethical and relevant. The framework is especially useful for checking whether a technically sound question is worth and able to be investigated.
| FINER test | Student-friendly question |
| Feasible | Can I answer this with the time, sources/data, access, skills and resources I actually have? |
| Interesting | Is the issue engaging enough to sustain careful research, and is it meaningful to the intended audience? |
| Novel | Does the project extend, compare, apply, clarify or revisit existing knowledge in a useful way? |
| Ethical | Could the research be conducted responsibly under the rules that apply to this assignment or study? |
| Relevant | Would answering the question contribute useful understanding to the course, field, practice or problem? |
“Novel” does not mean every undergraduate paper must discover something no one has ever studied. A student project can be worthwhile because it applies existing ideas to a selected case, synthesizes a debate, compares materials, or examines a bounded population. The standard should match the level and purpose of the assignment.
Step 9: Use PICO or PICOT Only When It Fits the Question
Clinical and intervention-focused questions often benefit from PICO or PICOT: population, intervention, comparison, outcome, and sometimes time. Harvard Catalyst’s research-question resources identify PICO as one useful framework for structuring clinical questions. It can make a question more searchable and explicit, but it is not designed for every humanities, qualitative or non-intervention project.
| Element | Planning prompt | Example |
| P – Population | Who is the question about? | Adults with hypertension |
| I – Intervention | What intervention or exposure is being examined? | Home blood-pressure monitoring |
| C – Comparison | What is it compared with, if relevant? | Usual care without home monitoring |
| O – Outcome | What result is being examined? | Blood-pressure control |
| T – Time | Over what period, if time is important? | Six months |
| Framework caution
Do not force a humanities interpretation, open-ended qualitative inquiry or non-intervention social-science question into PICO simply because the framework is familiar. Use a structure that matches the kind of evidence and answer your project actually needs. |
Step 10: Revise for Precision and Neutrality
Read the question word by word. Replace vague terms such as “things,” “effects,” “issues,” “good,” “bad,” or “successful” with concepts that can be defined. Remove unnecessary background phrases. Check whether the population, comparison or context is clear enough. Then check for assumptions hidden inside the wording.
A question such as “How does social media damage student wellbeing?” already assumes damage. A neutral version might ask how a particular pattern of use is associated with a defined wellbeing measure in a specified student group. Neutral wording does not mean the topic lacks importance; it means the evidence is allowed to determine the answer.
Step 11: Decide Whether You Need Sub-Questions
A substantial dissertation or multi-part project may need one main research question supported by a small number of sub-questions. Sub-questions should divide the central inquiry into logical parts, not introduce unrelated projects. A short course paper often works better with one strong question than with a long list.
| Main question | Possible sub-questions |
| How do first-year composition students perceive generative-AI feedback during revision? | What benefits do students report? / What limitations or concerns do they identify? / How do they decide whether to use a suggestion? |
| How do two national newspapers frame responsibility in climate-change coverage during a selected period? | How often is responsibility assigned to individuals? / How is institutional responsibility described? / What recurring frames distinguish the outlets? |
If each sub-question requires a different population, dataset or methodology, the project may be fragmented. Revisit the main question and remove branches that do not directly help answer it.
Research Question Examples Across Disciplines
The examples below illustrate structure rather than provide ready-made topics. A real question must be adapted to the assignment, evidence base, location, ethics requirements and level of study.
| Area | Example research question | Main question type |
| Education | How do first-year composition students perceive the usefulness and limitations of generative-AI feedback during revision? | Exploratory / qualitative |
| Nursing | What is the relationship between perceived workload and burnout among early-career medical-surgical nurses? | Relational |
| Business | How does meeting load differ between employees reporting high and low levels of focused work in hybrid professional teams? | Comparative |
| Psychology | How is late-night social-media use associated with self-reported sleep quality among first-year university students? | Relational |
| History | How did two selected newspapers frame women’s participation in a defined political campaign? | Comparative / interpretive |
| Literature | How does a selected novel use domestic space to represent social constraint? | Interpretive |
| Public health | Among adults with hypertension, does home blood-pressure monitoring compared with usual care improve blood-pressure control over six months? | Intervention / comparative |
Research Question vs. Hypothesis vs. Aims and Objectives
These elements are connected, but they should not be treated as interchangeable. The research question asks what the project will investigate. A hypothesis states a testable expected result when the design requires one. An aim states the overall purpose of the study, while objectives break that purpose into specific tasks or outcomes.
| Element | Function | Illustrative wording |
| Research question | Asks what the study seeks to answer. | What is the relationship between perceived workload and burnout among early-career medical-surgical nurses? |
| Hypothesis | Predicts an expected relationship or difference. | Higher perceived workload will be associated with higher burnout scores. |
| Aim | States the overall purpose. | To examine the relationship between perceived workload and burnout among early-career medical-surgical nurses. |
| Objective | States a specific action needed to achieve the aim. | To measure perceived workload and burnout using the selected instruments and test their association. |
Later EssayEco clusters on how to develop a research hypothesis and how to write research aims and objectives will cover those elements in detail. Keeping the distinctions clear here prevents the research question from becoming a prediction or a list of tasks.
Common Research Question Mistakes
- Writing a subject or title instead of a question.
- Using a yes/no question when the assignment requires analysis and synthesis.
- Combining several unrelated questions with repeated “and” clauses.
- Using vague terms such as “impact,” “success,” “better,” or “wellbeing” without defining what they mean in the project.
- Assuming the answer in the wording, such as asking why a policy “failed” before establishing that it failed.
- Claiming causation when the planned evidence can show only association or comparison.
- Choosing a population, dataset or source base that cannot realistically be accessed.
- Using PICO/PICOT for a question that is not clinical or intervention-oriented.
- Adding too many sub-questions to make the project look more sophisticated.
- Refusing to revise the question after preliminary research changes the terminology or feasible scope.
Research Question Checklist Before You Move On
| Check | Question to ask |
| Topic fit | Does the question emerge directly from a workable, assignment-appropriate topic? |
| Clarity | Are the key concepts understandable without extra explanation? |
| Focus | Can the question be answered in the available word count, time and project scale? |
| Evidence | Can realistic sources, texts, participants or data answer it? |
| Analysis | Does the question require more than a quick factual lookup? |
| Neutrality | Have I removed assumptions and loaded wording? |
| Design fit | Does the question ask for a kind of answer the likely research design can produce? |
| Feasibility | Do access, ethics, skills and resources make the project realistic? |
| Relevance | Can I explain why answering the question matters? |
| Wording | Is the sentence concise enough to keep the investigation visible? |
Frequently Asked Questions
What is the easiest way to learn how to write a research question?
Use a repeatable sequence: begin with a focused topic, read enough scholarship to understand the issue, identify what needs explaining or examining, choose the kind of answer you need, add necessary boundaries, draft several versions, test feasibility, and revise for clarity and neutrality. The question should become more precise as your understanding improves.
Can a research question change after I start researching?
Yes. A working question often changes when preliminary reading reveals better terminology, an inaccessible dataset, a more useful comparison, or a narrower gap. Revision is productive when it improves coherence and feasibility. Major late changes can create problems if the method, evidence collection or literature review has already been built around the earlier question.
Should a research question start with how or why?
Not necessarily. “How” and “why” often produce analytical questions, but “what,” “which,” “to what extent,” and direct comparison structures can also be appropriate. The opening should match the type of answer the project seeks rather than a formula.
How many research questions should a paper have?
There is no universal number. A short research paper may need one central question. A dissertation or complex project may use one main question and several tightly related sub-questions. The correct number depends on the discipline, assignment and design. Every question added should have a clear role in the same project.
Is a research question the same as a thesis statement?
No. The research question asks what the investigation seeks to answer. The thesis states the paper’s eventual claim or answer after sufficient research and analysis. In some assignments, the question remains implicit in the final paper while the thesis becomes the visible controlling statement.
Can a research question be a yes-or-no question?
Sometimes a formal scientific study may test a binary proposition, but for most research-based university writing, a simple yes-or-no structure is usually too limited. George Mason’s guidance recommends questions that require synthesis and analysis rather than an easily located factual answer.
