- A research question is a single, focused, answerable question that drives your entire study design, data collection, and analysis.
- Strong questions are specific, researchable, and open-ended; they name a population, a variable or concept, and a context.
- The FINER criteria (Feasible, Interesting, Novel, Ethical, Relevant) give you a 5-point test before you commit to a topic.
- Expect 3-5 revision rounds. Almost every first draft is too broad, too vague, or too large for the time available.
What Is a Research Question?
A research question is a clear, focused, and answerable question that a study sets out to investigate. It converts a broad topic into a specific inquiry that can be addressed with evidence, and it dictates your methods, sample, and analysis.
Every element of a project follows from it. Choose a weak question and you’ll have a diluted theoretical framework as well as hypotheses that you can’t test. No amount of good data collection will rescue the study.
A complete research question typically contains 3 components:
- Population or subject: who or what is being studied.
- Focus: the variable, experience, relationship, or intervention of interest.
- Context: the setting, time frame, or condition that bounds the study.
Research Question vs. Research Problem
A research problem is the broad issue or gap that makes a study worth doing; a research question is the specific, answerable question you ask about that problem. The problem justifies the work, and the question directs it.
| Aspect | Research Problem | Research Question |
| Form | 2-4 declarative sentences | 1 interrogative sentence, usually 10-30 words |
| Purpose | Establishes why the study matters | Defines exactly what the study will answer |
| Scope | Broad; can generate 10 or more questions | Narrow; addresses 1 specific angle |
| Where it appears | Introduction and problem statement | End of the introduction; repeated in methods and discussion |
| Test of quality | Does a reader accept that this gap matters? | Can this be answered with evidence you can collect? |
Worked example
- Problem: Nurse turnover in urban public hospitals has risen sharply since 2020, straining staffing budgets and patient safety, yet hospitals lack evidence on which retention measures actually work.
- Question 1: Which workplace factors predict 12-month turnover among nurses in urban public hospitals?
- Question 2: How do nurses who resigned within 12 months describe their decision to leave?
How they connect
The problem comes first and is written in declarative prose; the question is derived from it and must be narrow enough to answer within your time, data, and budget. If your question restates the problem word for word, it is almost certainly too broad. If you can answer the question but no reader sees why it matters, the problem statement is doing too little work.
Why Does a Strong Research Question Matter?
A strong research question matters because it determines feasibility, method, and the value of your findings. Reviewers judge the question first: if it is unclear or unanswerable, the rest of the proposal rarely survives scrutiny.
A well-built question delivers 5 practical benefits:
- It sets clear boundaries, so you know which literature to read and which to skip.
- It points to a method, because different question types demand different designs.
- It defines your data needs, including sample size, instruments, and access permissions.
- It keeps writing on track, giving every chapter a single reference point.
- It makes contribution visible, showing reviewers exactly what gap you fill.
Research Question vs Hypothesis vs Objective vs Problem Statement
A research question asks, a hypothesis predicts, objectives state what you will do, and a problem statement explains why the study is needed. All 4 describe the same project at different levels, so mixing them up is a frequent error in early drafts.
| Element | Form | Example |
| Problem statement | Declarative; states the gap and its consequence | Burnout among software engineers is rising, yet the role of remote work arrangements remains poorly understood. |
| Research question | Interrogative; open-ended and answerable | How does remote work affect burnout among software engineers in mid-sized firms? |
| Research objectives | Infinitive; action-based and numbered | 1. To measure weekly remote days. 2. To assess burnout scores. 3. To test the association between them. |
| Hypothesis | Declarative; testable prediction | Engineers working remotely 4 or more days per week report higher burnout than office-based peers. |
A simple sequence links all 4: the problem statement justifies the question, the question generates the objectives, and the objectives guide the hypothesis you test. Qualitative studies typically stop at objectives and omit the hypothesis.
What Are the Main Types of Research Questions?
There are 7 common types, and each one implies a different design. Descriptive and exploratory questions suit qualitative work; comparative, correlational, and causal questions usually require quantitative data.
| Type | What It Asks | Example |
| Descriptive | What is happening? | What study strategies do first-year medical students report using? |
| Exploratory | What is going on here, and why? | How do refugee parents describe their experience of school enrollment? |
| Comparative | Do 2 or more groups differ? | Do public and private hospitals differ in 30-day readmission rates? |
| Correlational | Are 2 variables related? | What is the relationship between screen time and sleep duration in teenagers? |
| Causal | Does X cause a change in Y? | Does a 4-day work week reduce voluntary turnover in retail firms? |
| Evaluative | How well does something work? | How effective is a peer-mentoring program at improving retention? |
| Predictive | What is likely to happen? | Which admission variables best predict first-year attrition? |
What Makes a Research Question Strong?
A strong research question is focused, complex, and researchable within your constraints. It cannot be answered with a simple yes, no, or a single fact lookup, and it can be answered with evidence you are able to collect.
| Quality | What It Means | Failure Signal |
| Clear | A reader understands it without extra explanation. | Uses jargon or undefined abstractions. |
| Focused | Narrow enough to answer thoroughly in the space available. | Could fill 5 dissertations. |
| Complex | Requires analysis and synthesis, not a lookup. | Answerable with 1 statistic. |
| Researchable | Evidence exists or can be gathered ethically. | Relies on inaccessible data. |
| Arguable | Reasonable people could reach different conclusions. | The answer is already obvious. |
| Relevant | Connects to a real gap or practical need. | No stakeholder cares about the answer. |
What Are the FINER Criteria for a Research Question?
FINER stands for Feasible, Interesting, Novel, Ethical, and Relevant. Introduced by Hulley and colleagues in clinical research design, it is now a standard 5-point screen used to test whether a question is worth pursuing.
Run each candidate question through all 5 criteria before you commit. A question that fails on feasibility or ethics should be redesigned or dropped, no matter how interesting it seems.
| Criterion | Core Question | Quick Test |
| Feasible | Can you actually do this? | Do you have the sample, time, skills, budget, and access? |
| Interesting | Does the answer intrigue you and your field? | Would 3 colleagues want to hear the result? |
| Novel | Does it add something new? | Has a systematic search shown the gap is real? |
| Ethical | Can it be done without harm? | Would an institutional review board approve it? |
| Relevant | Does the answer matter? | Will it change practice, policy, or theory? |
How Do You Apply FINER in Practice?
Score each criterion from 1 to 5, then compare candidates. This simple scoring converts a vague hunch into a decision you can defend to a supervisor or funding committee.
- Feasible: estimate participant numbers, then halve them. Recruitment almost always underperforms.
- Interesting: pitch the question in 30 seconds and watch whether people ask a follow-up question.
- Novel: search at least 2 databases plus a registry of ongoing trials or projects.
- Ethical: identify vulnerable groups, consent needs, and data protection duties before you draft.
- Relevant: name 2 specific audiences who would act on the finding.
What Are the Limits of the FINER Criteria?
FINER screens quality but does not generate ideas. It also weights novelty heavily, which can undervalue replication studies that the evidence base badly needs.
- It gives no guidance on wording or structure; pair it with PICO or SPIDER.
- Novelty is judged against published work, so unpublished null results stay invisible.
- Feasibility judgments are subjective; ask 2 people who have run similar projects.
Which Frameworks Help You Structure the Question?
Use PICO for quantitative and clinical questions, and SPIDER for qualitative or mixed-methods work. Both frameworks force you to name each component, which exposes the vague parts of a draft in minutes.
| Framework | Components | Best For |
| PICO | Population, Intervention, Comparison, Outcome | Clinical trials, effectiveness questions |
| PICOT | PICO plus Time frame | Studies with a defined follow-up period |
| SPIDER | Sample, Phenomenon of Interest, Design, Evaluation, Research type | Qualitative and mixed-methods studies |
| PEO | Population, Exposure, Outcome | Observational and epidemiological work |
| SPICE | Setting, Perspective, Intervention, Comparison, Evaluation | Policy, management, and service evaluation |
How Do You Write a Research Question Step by Step?
Follow 8 steps: pick a topic, read widely, find a gap, name your variables, draft several versions, test them against FINER, refine the wording, and confirm feasibility with a supervisor.
| Step | Action | Output |
| 1 | Choose a broad topic you can sustain for months. | Topic area |
| 2 | Read 15-20 recent reviews and key papers. | Reading notes |
| 3 | Identify a gap, contradiction, or untested context. | Stated problem |
| 4 | Name the population, focus, and context precisely. | 3 components |
| 5 | Draft 3-5 candidate questions in different forms. | Candidate list |
| 6 | Score each candidate against the FINER criteria. | Ranked shortlist |
| 7 | Rewrite the leader using precise, measurable wording. | Refined question |
| 8 | Check data access and timeline with your supervisor. | Approved question |
How Do You Narrow a Topic Into a Question?
Narrow by adding constraints one at a time: population, place, period, and outcome. Each constraint you add makes the study smaller, sharper, and far more likely to finish on schedule.
- Start: social media and mental health.
- Add population: social media and mental health among adolescents.
- Add focus: social media use and anxiety symptoms among adolescents.
- Add context: among adolescents aged 13-16 in urban public schools.
- Final: How is daily social media use associated with anxiety symptoms among adolescents aged 13-16 in urban public schools?
Examples of Good Research Questions
Good research questions name a population, a focus, and a context in 1 sentence. The 8 examples below span disciplines and show how the same structural rules apply across quantitative and qualitative work.
| Field | Research Question | Type |
| Public health | How does a 12-week walking program affect blood pressure in adults over 60? | Causal |
| Education | How do teachers describe their use of formative assessment in mixed-ability classrooms? | Exploratory |
| Psychology | What is the relationship between sleep duration and working memory in university students? | Correlational |
| Business | Do firms with employee stock ownership report lower turnover than comparable firms? | Comparative |
| Sociology | How do first-generation graduates describe belonging during their first year at university? | Exploratory |
| Environment | How has urban tree cover changed in Mumbai between 2010 and 2024? | Descriptive |
| Computer science | Which features best predict user churn in subscription mobile applications? | Predictive |
| Nursing | How effective is bedside handover at reducing medication errors on surgical wards? | Evaluative |
How Do You Fix a Weak Research Question?
Fix a weak question by diagnosing the specific fault, then rewriting to add precision. The 5 examples below show the most frequent problems and a stronger version of each.
| Weak Version | Problem | Stronger Version |
| Is social media bad? | Vague; value-laden; yes or no. | How is daily social media use associated with anxiety symptoms in adolescents aged 13-16? |
| What is climate change? | Answerable by a textbook. | How have shifting monsoon patterns affected smallholder rice yields in Maharashtra since 2015? |
| Why is education important? | Not arguable; no boundaries. | How does 1 additional year of secondary schooling affect lifetime earnings in low-income districts? |
| Does therapy work? | Undefined terms; no population. | How effective is 8-week group cognitive behavioral therapy at reducing relapse in adults with depression? |
| How can we fix homelessness? | Too broad; solution-seeking. | What factors predict 12-month housing retention among adults leaving supported accommodation? |
Common Mistakes
The most common mistakes are asking a yes or no question, packing 3 questions into 1 sentence, and choosing a scope you cannot finish. Each is easy to spot once you know the pattern.
- Yes or no phrasing: replace with how, what, or to what extent.
- Double-barreled questions: split any question containing “and” into 2 separate questions.
- Loaded assumptions: avoid wording that presumes the answer, such as why is X harmful.
- Undefined terms: define words like effective, quality, or success before you use them.
- Solution-seeking: “how can we fix X” is a policy brief, not an empirical study.
- Unavailable data: never build a question on records you have not confirmed you can access.
- Excessive scope: a global, multi-site question rarely fits a 1-year timeline.
- Method-first thinking: choose the study design after the question, not before.
How Many Research Questions Should You Have?
Most projects need 1 main question plus 2-4 sub-questions. Each sub-question should address a distinct component of the main question and map to a specific chapter, data set, or analysis.
| Project Type | Main Questions | Sub-Questions |
| Undergraduate essay | 1 | 0-2 |
| Master’s dissertation | 1 | 2-3 |
| Doctoral thesis | 1 | 3-5 |
| Journal article | 1 | 0-2 |
| Grant proposal | 1 per aim | 2-3 per aim |
How Research Questions Influence Study Design and Sampling
A research question dictates study design and sampling because its verb and scope define what evidence would count as an answer: causal wording (“does X reduce Y”) requires a comparison group and randomized or probability-based sampling, while exploratory wording (“how do people describe X”) requires qualitative methods and purposive sampling of information-rich cases.
| Question Type | Example Question | Typical Study Design | Sampling Method |
| Descriptive | What study strategies do first-year medical students report using? | Cross-sectional survey; observational audit | Simple random or stratified random sampling |
| Exploratory | How do refugee parents describe enrolling a child in school? | Qualitative: interviews, focus groups, grounded theory | Purposive sampling; snowball sampling to reach hidden groups |
| Comparative | Do public and private hospitals differ in 30-day readmission rates? | Cross-sectional or cohort comparison; quasi-experimental | Stratified sampling by group; matched sampling on key covariates |
| Correlational | What is the relationship between screen time and sleep duration in teenagers? | Cross-sectional survey; correlational or panel design | Probability sampling, ideally random or cluster, to protect generalizability |
| Causal | Does a 4-day work week reduce voluntary turnover in retail firms? | Randomized controlled trial; difference-in-differences; interrupted time-series | Random assignment after recruitment; cluster randomization for team-level treatments |
| Evaluative | How effective is peer mentoring at improving student retention? | Program evaluation; pre-post design with comparison group | Census of program participants plus a matched comparison group |
| Predictive | Which admission variables best predict first-year attrition? | Retrospective cohort; predictive modeling on existing records | Full census of administrative records; split into training and holdout samples |
| Mixed-methods | How and why does a nutrition program change eating behavior? | Explanatory sequential or convergent design | Probability sampling for the survey phase; purposive subsample for interviews |
Three practical notes on reading this table:
Sample size follows the design, not the question. Causal and correlational questions need power calculations before recruitment; exploratory questions instead target saturation, usually reached at 12-20 interviews.
Access constraints often override the ideal method. If random sampling is impossible, say so explicitly and use convenience or quota sampling, then discuss the limitation rather than hiding it.
Question wording is the tell. Verbs like “cause,” “reduce,” or “improve” commit you to a design with a comparison group. If you cannot build one, rewrite the question as correlational before you collect data.
Checklist for Finalizing a Research Question
Run your question through these 14 checks. If you cannot answer yes to every item, revise before writing your proposal, because each unresolved item creates a problem later in the project.
Is the Question Clear and Focused?
- Does it fit in 1 sentence, ideally under 30 words?
- Does it name a specific population or subject?
- Does it name a specific focus, variable, or phenomenon?
- Does it state a context, setting, or time frame?
- Are all key terms defined or measurable?
Is the Question Answerable?
- Does it start with how, what, why, to what extent, or a similar open form?
- Does it require analysis rather than a single fact?
- Do you know which data would answer it?
- Can you access that data within your timeline and budget?
- Does 1 clear method follow from the wording?
Is the Question Worth Asking?
- Does a literature search confirm the gap is real?
- Would the finding change practice, policy, or theory?
- Does the question pass all 5 FINER criteria?
- Has a supervisor or peer confirmed it is realistic?
Frequently Asked Questions
What Is a Good Research Question Example for a Dissertation?
A good example is: How does 1 additional hour of daily remote work affect self-reported burnout among software engineers in mid-sized firms? It names a population, a measurable focus, and a context, and it can be answered with survey or panel data in 1 academic year.
How Long Should a Research Question Be?
Aim for 1 sentence of 10-30 words. If your draft runs longer, you are probably asking 2 questions at once; split it into a main question and 1 or more sub-questions instead.
What Is the Difference Between a Research Question and a Research Problem?
The research problem is the broader issue that needs attention; the research question is the specific, answerable question you ask about it. 1 problem can generate 10 or more distinct questions.
Can a Research Question Be a Yes or No Question?
Generally no. Closed questions produce thin answers and weak analysis. Convert them by asking to what extent, how, or under what conditions, which invites evidence, nuance, and comparison.
How Do You Write a Research Question for a Qualitative Study?
Use how or what, focus on experience or meaning, and avoid variables and causal language. Example: “How do first-generation students describe belonging during their first year at university?” Use the SPIDER framework to structure it.
How Do You Turn a Research Topic Into a Research Question?
Add constraints one at a time. Start with the topic, then add a population, a specific focus, and a context or time frame. After 3-4 constraints, most topics become a workable question.
What Are the 5 FINER Criteria for a Good Research Question?
They are Feasible, Interesting, Novel, Ethical, and Relevant. Score each candidate question from 1 to 5 on every criterion, then choose the highest scorer that you can realistically complete.
How Many Research Questions Should a Master’s Dissertation Have?
Use 1 main question and 2-3 sub-questions. More than 4 questions usually signals that the scope is too wide for a project of 12,000-20,000 words and a single data collection round.
Important Definitions
| Term | Definition |
| Research question | The specific, answerable question a study is designed to address. |
| Research problem | The broader issue, gap, or difficulty that makes the question worth asking. |
| Research aim | A statement of overall purpose; usually 1 aim covers 2-4 questions. |
| Hypothesis | A testable, predictive statement about the relationship between variables. |
| Variable | Any characteristic that can change or differ across cases, such as age or income. |
| Independent variable | The factor you manipulate, compare, or treat as the presumed cause. |
| Dependent variable | The outcome you measure and expect to be influenced. |
| Population | The full group your findings are meant to describe, such as first-year nurses. |
| Scope | The boundaries of the study: who, what, where, and over what period. |
| Operational definition | A precise statement of how an abstract concept will be measured. |
| Research gap | A question that existing literature has not answered, or has answered poorly. |
| Feasibility | Whether the study can be completed with your time, data, skills, and budget. |
Comment