How to Write a Research Question Step by Step
From a broad topic to a focused, answerable inquiry — the complete process for forming research questions across every discipline, with examples, evaluation frameworks, and the most common formulation errors explained.
Every research paper, dissertation, and academic essay begins with the same problem: something you want to understand, and the task of turning that interest into a question specific enough to answer. The gap between “I am interested in climate change and economic inequality” and a research question that can actually drive a study is where most students stall. The topic is interesting; the question remains formless. This guide works through that transformation systematically — from broad interest to focused inquiry, from intuition to a question a methodology can answer. Every step is illustrated with examples drawn from real disciplines. By the end, you will be able to formulate, evaluate, and refine a research question for any paper, at any level, in any field.
What a Research Question Is — and What It Has to Do
A research question is a focused, specific, answerable inquiry that defines what a study is trying to find out. It is not a topic, not a title, and not a problem statement — though all three contribute to forming it. A topic is a broad area of interest. A problem statement describes a situation requiring investigation. A research question specifies exactly what you will investigate about that situation, in what population, in what context, and in what direction.
The research question does more structural work in a paper than any other single element. It determines what counts as relevant literature in your review. It governs which methodology can answer it. It defines which data needs to be collected. It sets the scope of the analysis. And it provides the standard against which your conclusions are evaluated: a conclusion is only convincing if it answers the question you set out to answer. A vague, broad, or poorly formed question produces a paper that cannot be coherently organised, a literature review that sprawls without direction, and findings that do not add up to a clear answer. The time invested in forming a precise question at the start of a project saves many more hours downstream.
A research question has two equally important properties that pull in opposite directions: it must be specific enough to be answerable by a study of finite scope, and it must be significant enough that answering it matters. Too narrow and you have a trivial question with a trivial answer. Too broad and you have a topic, not a question. The entire skill of research question formation is finding the productive middle ground — the point where focus and significance coincide. The frameworks and steps in this guide are practical tools for locating that point.
The Five Types of Research Questions
Research questions are not a single genre. They vary by what kind of answer they are seeking — and that variation directly determines the methodology required to answer them. Identifying which type of question you are asking is one of the first steps in the formulation process, because it clarifies both the scope of the inquiry and the research design it demands.
What Is the Current State of X?
Descriptive questions ask what exists, what is happening, or what characteristics something has. They do not seek to explain causes, test relationships, or evaluate outcomes — they document a phenomenon or population. Examples: “What are the sleep patterns of first-year university students during examination periods?” or “What proportion of small businesses in Kenya use mobile payment platforms for payroll?” Descriptive questions require observational, survey, or archival methodology. They are the foundation for subsequent explanatory and causal inquiry — you cannot test why something happens until you have established that it does.
How Does X Differ Between Groups or Conditions?
Comparative questions examine differences between two or more groups, settings, time periods, or conditions. Examples: “How do maternal mortality rates differ between rural and urban health facilities in sub-Saharan Africa?” or “How does student engagement with synchronous vs asynchronous online learning differ by prior academic performance?” Comparative questions require at least two clearly defined categories of comparison and a specified outcome measure. Their methodology ranges from cross-sectional surveys to quasi-experimental designs depending on whether group assignment is natural or manipulated.
Does X Cause or Produce Y?
Causal questions investigate whether and how one variable produces change in another. They are the strongest form of research question in terms of the evidence they require — and the hardest to answer validly. Examples: “Does mindfulness-based stress reduction training reduce cortisol levels in adult patients with generalised anxiety disorder?” or “Does increasing the minimum wage by 10% reduce employment rates in low-skill sectors?” True causal answers require experimental designs (RCTs, quasi-experiments). Observational designs can establish association but not causation — a distinction your question framing and methodology section must acknowledge.
What Is the Relationship Between X and Y?
Relationship questions investigate whether two or more variables are associated, how strongly, and in what direction — without asserting causation. Examples: “What is the relationship between social media use frequency and loneliness scores in adults aged 60 and over?” or “Is there a significant correlation between class attendance rates and final examination performance in undergraduate economics courses?” Relationship questions are answered through correlational and regression analyses. They are more modest than causal claims and more honest when experimental designs are not feasible — which is most of the time in social and educational research.
How Effective, Appropriate, or Successful Is X?
Evaluative questions assess whether something achieves its stated purpose, meets a defined standard, or produces intended outcomes. Examples: “How effective is the WHO’s HEARTS package in reducing cardiovascular disease risk in low-resource primary care settings?” or “To what extent does the UK’s Early Years Foundation Stage curriculum meet the developmental needs of children with English as an additional language?” Evaluative questions require defining explicit criteria for evaluation before data collection, not after. They are common in policy analysis, programme evaluation, and professional practice disciplines.
What Factors, Experiences, or Meanings Underlie X?
Exploratory questions are primarily qualitative — they seek to understand a phenomenon in depth, to identify themes and factors not yet well-documented, or to understand the lived experience of a population. Examples: “What factors influence the decision of rural Kenyan women to seek formal antenatal care?” or “How do first-generation university students construct their academic identity in the first year of study?” Exploratory questions do not test hypotheses; they generate them. They are answered through interviews, focus groups, ethnography, and thematic analysis. They are appropriate when the literature on a topic is thin or when quantitative data exists but the mechanisms and meanings behind it are not yet understood.
From Topic to Focused Research Question: The Step-by-Step Process
Most students begin with a topic they find interesting — a broad area, a social issue, a disciplinary problem. The work of forming a research question is the work of progressive specification: narrowing, focusing, and ultimately defining the exact inquiry that a study can address. The process below applies to any discipline and any level of study.
Identify a broad topic area and state your initial interest
Begin with the broadest honest description of what you want to understand. Do not self-censor at this stage — the point is to articulate the interest before refining it. Write it as a statement: “I am interested in how poverty affects children’s educational outcomes.” This is not yet a research question. It is a starting point. The more specific your initial interest, the fewer narrowing steps you will need — but even a broad topic is a valid starting point if the narrowing process is applied systematically.
Conduct preliminary literature reading to identify what is already known
Before forming a specific question, you need to know what has already been investigated. Fifteen to twenty articles on your broad topic will reveal where the gaps, debates, and unanswered problems are — and those are the territories where a new research question lives. Look for recurring limitations sections (“future research should address…”), contested findings where different studies reach different conclusions, and populations or contexts underrepresented in existing work. Your question should emerge from a real gap in the literature, not from what you personally find interesting in isolation from what is already known.
Define the specific population you are studying
“Children” is not a population — it is a demographic category containing every human being under eighteen in every country in every circumstance. “Primary school children aged 8–11 in urban public schools in Lagos” is a population. Specifying your population answers the “who” of your question and immediately constrains the scope to something researchable. Consider: age range, geographic context, institutional setting, relevant demographic characteristics, and any inclusion or exclusion criteria that matter to the phenomenon you are investigating. Your population should be broad enough to generate meaningful conclusions and narrow enough to be practically studiable.
Identify the key variable or phenomenon of interest
What specifically about your topic are you investigating? “Educational outcomes” could mean test scores, dropout rates, literacy levels, university attendance, teacher assessments, or a dozen other things. “Economic hardship” could mean household income, food insecurity, housing instability, or parental employment status. Commit to specific, measurable (or at least clearly definable) variables. In quantitative research, distinguish between your independent variable (the cause or predictor) and your dependent variable (the outcome). In qualitative research, identify the phenomenon of interest — the experience, process, or meaning you are exploring.
Define the relationship type you are investigating
Using the question type framework from Section 2, determine what kind of relationship you are asking about: are you describing, comparing, explaining causation, measuring association, evaluating effectiveness, or exploring meaning? This determines the question’s grammatical structure and the methodology it requires. “How does X affect Y?” is causal. “Is there a relationship between X and Y?” is correlational. “What are the experiences of X?” is exploratory. The relationship type is not arbitrary — it should reflect both what the literature gap suggests is needed and what is methodologically achievable within your project’s scope.
Draft the question and check that it is answerable
Combine your population, key variable, relationship type, and context into a single interrogative sentence. Then apply the answerability test: can you identify a methodology that would produce data capable of answering this question? If the answer is “I am not sure what data would answer this,” the question is still too vague. If the answer is “it would take a twenty-year longitudinal study with unlimited funding,” the question is not feasible at your scale. A good draft question points clearly toward a study you could actually conduct — or, for a literature review or theoretical paper, toward evidence that exists and can be synthesised.
Apply the FINER criteria to evaluate and refine
Take your draft question through the FINER checklist (detailed in Section 4 below). This structured evaluation will identify specific weaknesses — a question that is interesting but not feasible, or relevant but not novel — and direct you toward targeted revisions rather than wholesale reformulation. Most draft questions pass three or four FINER criteria and need work on one or two. The checklist makes the problem visible and specific, which makes revision tractable rather than overwhelming.
Confirm alignment with your methodology and available resources
The final test before committing to a research question is practical: do you have access to the population your question specifies? Can you obtain ethical approval for the data collection it requires? Do you have the time and analytical skills to answer it within the project timeline? Does the question fit the word count and scope of the assignment? A dissertation research question has different scale requirements than a 3,000-word undergraduate essay. A question that requires primary data collection is differently constrained than one answerable through literature review and secondary analysis. Ensure your question fits your actual resources before building a methodology around it.
A Worked Example: From Topic to Finalized Research Question
The narrowing process is easier to understand through a concrete example. The following traces one topic through each step:
Evaluating Your Research Question: The FINER Criteria
The FINER criteria — developed in the context of clinical research methodology and subsequently adopted across social science, education, and business research — provide a five-point framework for evaluating whether a research question is ready to drive a study. Each criterion addresses a distinct risk: questions that fail on Feasibility waste resources on unanswerable investigations; questions that fail on Novelty produce work the field already has. Running every draft question through all five criteria before committing to it is one of the most efficient investments of research time available.
F — Feasible
Can this question be answered with available time, resources, population access, and skills? A feasible question at undergraduate level is different from a feasible question at doctoral level. Test: would the methodology required to answer this question be practicable within your project constraints?
I — Interesting
Is the question of genuine interest to you, your discipline, and the broader audience who might read the findings? Interesting does not mean trendy — it means the question’s answer would matter to someone beyond the researcher. Prolonged research on a question you do not find interesting produces poor-quality work.
N — Novel
Does the question address a gap, extend existing findings, replicate important work in a new context, or challenge received wisdom? A question whose answer is already well-established in the literature is not novel. Novel does not mean unprecedented — a replication in a different population or setting is novel if that difference matters.
E — Ethical
Can the question be investigated without causing harm to participants, violating privacy, or producing findings that could be misused? All research involving human subjects requires ethics consideration. A question is unethical if answering it requires deception, exposes vulnerable populations to risk, or produces data whose harm potential exceeds its knowledge benefit.
R — Relevant
Does the question advance scientific or scholarly knowledge in a meaningful direction? Is it relevant to current debates, pressing social problems, or the development of theory in its discipline? Relevant questions connect to something larger than the immediate study — they speak to an ongoing conversation in the field, address a problem with real-world implications, or test a theory that has broader applications. A question can be feasible, interesting, novel, and ethical and still fail the relevance criterion if its findings would not matter to anyone beyond the researcher.
The FINER criteria are diagnostic, not disqualifying. A question that fails on Feasibility may need scope reduction — a smaller sample, a different data source, or a narrower time frame — rather than complete abandonment. A question that fails on Novelty may need recontextualisation — the same question applied to an under-studied population or setting may be highly novel. A question that fails on Relevance may need a revised framing that connects it more explicitly to current theoretical debates or policy concerns.
The most common failure modes are Feasibility (students consistently overestimate what a single study can accomplish) and Novelty (students formulate questions whose answers are already well-established). Both are fixable with targeted revision rather than wholesale restart.
The PICO and PICOT Frameworks for Clinical and Health Sciences Research
In health sciences, nursing, medicine, and evidence-based practice, research questions are often formulated using the PICO framework — a structured approach that simultaneously forms the research question and generates the search terms for the systematic literature review required to answer it. PICO is not simply an acronym to memorise; it is a discipline-specific methodology for question formation that has become the standard across clinical research, nursing research, and EBP assignments globally.
Population
Who are you studying? Define by diagnosis, age, condition, setting, or other relevant characteristics. E.g., “adult patients with type 2 diabetes in primary care”
Intervention
What intervention, treatment, exposure, or factor are you examining? E.g., “dietary counselling delivered by a registered dietitian”
Comparison
What is the alternative? Standard care, placebo, another intervention, or no treatment. E.g., “standard care with no structured dietary guidance”
Outcome
What are you measuring? Specify primary and secondary outcomes. E.g., “HbA1c levels at 6 months” and “patient-reported quality of life scores”
Time (PICOT)
Over what time period is the outcome measured? E.g., “over a 12-month follow-up period.” Not always required but strengthens precision for longitudinal outcomes.
Assembled, a PICO question reads: “In adult patients with type 2 diabetes in primary care (P), does dietitian-delivered dietary counselling (I), compared with standard care without structured dietary guidance (C), reduce HbA1c levels at 6 months (O)?” This question is specific, answerable, points toward a clear systematic review or RCT methodology, and generates search terms directly from each PICO element.
When to Use PICO vs Other Frameworks
PICO is most appropriate for quantitative clinical questions about the effectiveness of interventions, treatments, or diagnostic approaches. The Comparison element requires that there is an identifiable alternative to compare against — which means PICO is less suitable for questions about the prevalence or incidence of conditions (which are descriptive), questions about patient experiences (which are qualitative and exploratory), or questions about the meaning of health-related phenomena.
For qualitative health research, the PICo framework (Population, phenomenon of Interest, Context) is more appropriate. For questions about prognosis and risk factors, PECO (Population, Exposure, Comparison, Outcome) replaces Intervention with Exposure. For diagnostic accuracy questions, PIRD (Population, Index test, Reference standard, Diagnosis) applies.
For nursing PICOT project assignments specifically, the NIH’s National Library of Medicine provides guidance on PICO question formulation as part of its evidence-based medicine resources, and most nursing programme syllabi include PICO templates as required assignment scaffolding. See our PICOT project writing support for assignment-specific guidance.
Qualitative vs Quantitative Research Questions — How They Differ in Form and Function
The distinction between qualitative and quantitative research questions is not just stylistic — it reflects fundamentally different philosophical orientations toward what knowledge is, how it is produced, and what counts as an answer. Understanding these differences prevents one of the most common research design errors: using a qualitative question structure with a quantitative methodology, or vice versa.
Mixed Methods Research Questions
Mixed methods studies combine quantitative and qualitative questions to answer different aspects of a research problem that no single methodology can fully address. A mixed methods research design typically has one overarching research question and two or more sub-questions — one quantitative and one qualitative — that together address the full scope of the inquiry.
Example: Overarching question: “How do school feeding programmes affect the nutritional outcomes and school engagement of primary school children in rural Ghana?” Quantitative sub-question: “Do children enrolled in the school feeding programme show significantly higher anthropometric measurements and school attendance rates than non-enrolled peers?” Qualitative sub-question: “How do parents, teachers, and children describe the role of the school feeding programme in children’s daily school experience?”
Mixed methods questions are most appropriate when a quantitative finding needs contextual explanation, when a qualitative finding needs to be tested at scale, or when the phenomenon under investigation has both objective measurable components and subjective experiential ones that are equally important. The integration rationale — why both methods are needed and how they relate — should be explicit in the research question structure itself.
Five Dimensions for Narrowing a Research Question That Is Too Broad
Broad research questions are the single most common formulation problem at every level of academic study. “How does stress affect health?” is not a research question — it is a subject area that has generated thousands of studies across multiple disciplines over several decades. The narrowing process is not about making your question less interesting; it is about making it answerable. A question that answers a small, specific thing well contributes more to knowledge than a question that gestures at everything and answers nothing.
Every broad question can be narrowed across five specific dimensions. Applying all five produces a question specific enough to drive a study; applying two or three is often sufficient to transform an unusable broad question into a workable focused one.
| Dimension | What to Specify | Broad Version | Narrowed Version |
|---|---|---|---|
| Population | Age, diagnosis, occupation, geographic location, institutional setting, and any defining inclusion/exclusion characteristics | “healthcare workers” | “registered nurses in ICU settings in South African public hospitals” |
| Variable / Phenomenon | Specific measurable variable (quantitative) or clearly defined phenomenon (qualitative); not a general category | “burnout” | “emotional exhaustion subscale of the Maslach Burnout Inventory” |
| Time Frame | The period over which the phenomenon is measured or to which the question applies | “during the pandemic” | “during the first 12 months of the COVID-19 pandemic (March 2020–March 2021)” |
| Context / Setting | The specific environment, institution, geographic region, or cultural context of the study | “in hospitals” | “in tertiary public hospitals in Johannesburg” |
| Relationship Type | Whether you are describing, comparing, explaining, correlating, evaluating, or exploring — which determines the question’s grammatical structure | “and job performance” | “and self-reported clinical decision-making quality” |
Applying all five dimensions to a starting broad question — “How does stress affect healthcare workers?” — produces: “What is the relationship between emotional exhaustion scores and self-reported clinical decision-making quality among registered nurses in ICU settings in Johannesburg tertiary hospitals during the first 12 months of the COVID-19 pandemic?” This is a researchable question. It specifies who, what, where, when, and what kind of relationship — which means it specifies the methodology, the data needed, and the analysis required.
Narrowing can go too far. A question so specific that only a handful of people in the world could be in the study population, or so granular that the findings would apply to no other context, is not a useful research question — it is a description of a single case. The target is a question specific enough to be answerable and broad enough to be meaningful.
A useful test: if you answered this question, who would care? If the answer is “basically no one outside the three people I studied,” the question is too narrow. If the answer is “anyone working in this field and this context, plus anyone studying related populations,” the specificity is about right. Significance and feasibility must both be present — not one at the expense of the other.
Research Question vs Hypothesis vs Thesis Statement
These three terms are frequently confused — used interchangeably in student writing when they are in fact distinct constructs that serve different functions in a research paper. Understanding the distinction clarifies where each belongs in a paper’s structure and prevents the errors that arise from conflating them.
Research Question
An open interrogative inquiry defining what the study aims to find out. It does not predict an outcome; it specifies the investigation. Present in all empirical papers, essays, and dissertations. Appears in the introduction. The paper exists to answer it.
Hypothesis
A specific, testable prediction about the expected relationship between variables — derived from theory or prior evidence. Used in quantitative research. States what the researcher predicts the data will show. Tested rather than answered. Can be supported or refuted by the findings.
Thesis Statement
A declarative statement of the central argument or conclusion a paper will defend. It answers the research question rather than asking it. In argumentative essays, it appears at the introduction’s end. In empirical papers, it emerges from the findings. Statements, not questions.
Research question: “Does sleep deprivation affect working memory performance in adults?”
An open question — specifies what is being investigated but does not predict the outcome. Defines the study without foreclosing its answer.
Hypothesis: “Adults who sleep fewer than six hours per night will demonstrate significantly lower working memory scores on the N-back task than those sleeping seven to nine hours.”
A testable prediction with direction, specificity, and measurement — derived from the research question and from prior sleep research. Can be supported or refuted by the data.
The relationship between these three is sequential: the research question defines the inquiry; the hypothesis (where applicable) predicts the answer; and the thesis statement (in argumentative or post-findings writing) states the conclusion. In a research paper, the research question appears in the introduction, the hypothesis in the methodology or introduction, and the thesis-equivalent conclusion emerges in the discussion. They are not alternatives — they occupy different positions in the paper’s architecture and do different work.
Research Questions and Research Objectives — How They Work Together
Research questions and research objectives are closely related but structurally distinct. A research question asks what the study will find out, framed interrogatively. A research objective states what the study will do, framed as an action, typically beginning with an infinitive verb. Together, they define both the inquiry and the means by which it will be conducted.
The four objectives above collectively describe the steps through which the primary research question will be answered. Each objective is specific, methodologically oriented, and contributes to a component of the answer. Together they constitute the study’s operational roadmap. Note that objectives use verbs like “to assess,” “to measure,” “to examine,” and “to identify” — active, methodologically grounded verbs rather than vague terms like “to investigate,” “to look at,” or “to explore” (which tell you nothing about the method).
Descriptive: to describe, to document, to characterise, to profile, to map, to catalogue.
Comparative: to compare, to contrast, to differentiate, to assess differences between.
Explanatory/causal: to determine, to test, to establish, to measure the effect of.
Evaluative: to evaluate, to assess, to appraise, to determine the effectiveness of.
Exploratory/qualitative: to explore, to understand, to identify themes in, to examine the meaning of, to examine how participants experience.
Research Question Examples Across Academic Disciplines
Seeing well-formed research questions across multiple disciplines clarifies how the same structural principles — specific population, defined variable or phenomenon, explicit relationship type, appropriate scope — manifest differently in different fields. The examples below are annotated to show what makes each one work.
Peer Victimisation and Academic Motivation
“Does experience of peer victimisation in secondary school predict lower intrinsic academic motivation at university entry, controlling for prior academic achievement?”
Causal-predictive. Specific population (secondary school students followed to university entry). Controlled design implied.
Urban Green Space and Mental Health
“What is the association between proximity to urban green space and self-reported anxiety and depression scores among adults in high-density residential areas in Nairobi?”
Correlational. Geographically specific. Both variables are operationally defined. Does not overclaim causation.
Teacher Feedback and Writing Quality
“How do Year 9 students in Kenyan public secondary schools describe the effect of written teacher feedback on their motivation to revise written assignments?”
Qualitative/exploratory. Population is specific. Focuses on student perception and meaning, not objective measurement.
Remote Work and Team Cohesion
“To what extent does the proportion of time spent in remote work settings predict team cohesion scores among project teams in Kenyan financial services firms?”
Predictive relationship. Sector-specific. Both IV and DV measurable. Industry context anchors the question’s relevance.
Mandatory Sentencing and Recidivism
“Do mandatory minimum sentencing provisions for drug offences reduce recidivism rates among first-time drug offenders in sub-Saharan African jurisdictions?”
Causal evaluative. Specific legal mechanism. Outcome (recidivism) is measurable. Geographic scope clarified.
Nurse-Patient Ratios and Adverse Events
“In adult acute-care hospital wards, does a nurse-to-patient ratio below 1:4 increase the incidence of adverse events compared with ratios of 1:4 or better?”
PICO-structured. Threshold specified. Comparison is explicit. Outcome (adverse events) is defined and measurable.
Environmental Science
“How does land-use change from subsistence farming to commercial floriculture affect soil microbial diversity in the Rift Valley highlands of Kenya over a 10-year period?”
History / Humanities
“How did colonial land tenure legislation in Kenya between 1895 and 1963 shape patterns of smallholder land access that persist in contemporary land disputes?”
Computer Science / IS
“How effectively does federated learning preserve data privacy compared with centralised machine learning in clinical prediction models for sepsis risk in ICU patients?”
Research Questions for Dissertations and Theses
A dissertation or thesis research question carries a different weight than the research question for a course essay. It defines a project that will consume months or years, must demonstrate an original contribution to knowledge (at postgraduate and doctoral levels), must be approved by a supervisor and sometimes an ethics board, and must be answerable within the constraints of a single-researcher project conducted without the resources of a funded research team. These additional demands shape how dissertation research questions are formed, evaluated, and refined.
The primary research question should be singular and central — all sub-questions answer components of it
A dissertation with five primary questions of equal weight is not a dissertation — it is five separate studies. The primary question should be the one whose answer constitutes the dissertation’s central contribution. Sub-questions break it into investigable components. If you cannot identify which of your questions is primary, the conceptual architecture of the project needs clarification before the research design can proceed.
Strong Dissertation Question Characteristics
Answerable within the project scope. Addresses a demonstrable gap. Specifies population, phenomenon, and context. Implies a methodology. Original enough to represent a genuine contribution at the relevant level.
Questions That Need Revision
Too broad to answer in one project. No clear gap identified in the literature. Population not specified. Answerable by a simple database search. Interesting to the student but not to the discipline.
Questions That Need Abandonment
Answerable with yes or no. Asks what is already conclusively established. Requires data that cannot be ethically or practically obtained. Requires a research team, funding, or timeline beyond the project’s scope.
At doctoral level, the research question must additionally demonstrate that it addresses a genuine gap in the existing literature — not just a topic the student finds interesting, but a specific unanswered question whose answer would advance the field. This requires a comprehensive literature review before the question is finalised, not after. Many doctoral students arrive at their final research question through the literature review process: they begin with a broad interest, conduct systematic literature reading, identify the specific gap or unresolved debate, and form the question around that gap rather than around the initial interest. The question that emerges from this process is almost always more intellectually rigorous than the question the student would have started with. For comprehensive dissertation support, see our dissertation writing service.
When Your Research Question and Methodology Do Not Align
The most consequential research design error is a mismatch between the question and the methodology. A causal question answered by a correlational design. A qualitative exploration question answered by a survey. An evaluative question applied to a study that collected no outcome data. These mismatches are almost always traceable to a research question that was not precise enough to specify the methodology it required. If your methodology chapter is difficult to write, revisit the research question first — the problem is usually there.
Common Research Question Formulation Errors
The errors below appear consistently across undergraduate and postgraduate writing. They are not random mistakes — each reflects a specific misunderstanding of what a research question is supposed to accomplish. Identifying which error your draft question makes points directly to the revision needed.
Relative frequency of research question formulation errors in undergraduate and postgraduate academic writing, based on recurring patterns identified in research methodology and academic writing pedagogy literature.
The Yes/No Problem Explained
A research question that can be answered with “yes” or “no” is not a research question — it is a factual check. “Does exercise reduce blood pressure?” can be answered: yes. The answer is established in the literature. What drives genuine research is the follow-up: by how much, in which populations, under what conditions, through what mechanisms, and compared with what alternatives? The yes/no question forecloses all of that. Convert it: instead of “Does X affect Y?” ask “How does X affect Y?” or “To what extent does X affect Y in population Z?” The revision opens the question to a range of possible answers rather than a binary.
Refining a Draft Research Question: A Practical Checklist
A draft research question is never the final research question. The refinement process — running the draft through a systematic checklist, testing it against the methodology it implies, and revising based on what you find — typically produces a substantially better question with each pass. The checklist below can be applied to any draft question at any level of study.
Is it a question? Does it end with a question mark and invite a range of possible answers?
If it is a statement, convert it. If it can only be answered with yes or no, add “to what extent,” “how,” or “in what ways” to open it to a range. A question that only one answer can answer is not a question — it is a conclusion stated interrogatively.
Is the population clearly specified? Can you identify who is in the study and who is not?
If the question could apply to “everyone” or “anyone,” the population is not specified. Narrow by age, geography, diagnosis, institutional setting, or any combination of characteristics that defines the group whose experience or data is relevant to answering the question.
Are the variables or phenomena of interest defined? Are they measurable or clearly describable?
Abstract terms like “wellbeing,” “engagement,” “success,” or “impact” need operational definition — what specifically will you measure or describe? In quantitative research, specify the measurement instrument or indicator. In qualitative research, specify the aspect of experience or phenomenon you are exploring.
Does the question’s structure match its intended methodology?
“What are the lived experiences of…” requires qualitative methodology. “To what extent does X predict Y…” requires quantitative regression analysis. “Does intervention X reduce outcome Y compared with Z…” requires a comparative or experimental design. If the methodology you are planning does not match the question’s grammatical structure, one of them needs to change.
Is the answer not already in the literature? Does the question address a genuine gap?
Search your exact question in Google Scholar and your institutional databases. If the first page of results contains three studies that appear to answer your question directly, your question is not novel. You either need to replicate in a different population, extend in a different context, or reformulate around a genuinely unaddressed aspect of the topic.
Does it pass all five FINER criteria? Identify which criterion (if any) is weakest and target the revision there.
Most questions pass FINER on three or four dimensions — the revision work is concentrated on the one or two where the question is weakest. A question that fails Feasibility needs scope reduction. One that fails Novelty needs recontextualisation or a literature gap it genuinely addresses. One that fails Relevance needs a clearer connection to current debates or practical significance in the field.
The Scribbr guide to research questions provides additional examples and a complementary checklist that is particularly useful for students working on quantitative social science research. For guidance on how a well-formed research question anchors the full paper, see the Purdue Online Writing Lab’s research paper writing guide. For clinical research using the PICO framework specifically, the NCBI’s evidence-based medicine resources include guidance on formulating focused clinical questions and translating them into systematic search strategies.
Research Questions for Literature Reviews, Systematic Reviews, and Secondary Research
Not all research questions drive primary data collection. Many undergraduate and postgraduate papers are literature reviews, systematic reviews, or secondary analyses — and the research questions for these papers have a distinct character from those driving primary empirical studies. The question still needs to be specific, answerable, and novel, but “answerable” here means answerable through synthesis of existing literature rather than through original data collection.
Literature Review Research Questions
A literature review research question asks what the existing body of evidence shows about a topic — not what a new study would find, but what the accumulated scholarship has established. Example: “What does the literature from 2010 to 2024 show about the relationship between teacher feedback practices and student self-regulation in secondary education?” This question is answerable by synthesising the existing literature without collecting new data. It must still be specific (defined time range, defined population, defined variables), novel (not already systematically reviewed), and relevant. For academic literature review writing support, see literature review writing services.
Systematic Review Research Questions
Systematic review questions follow the PICO or similar frameworks and are formulated with the same level of specificity as a primary study question — because the inclusion and exclusion criteria for the review are derived directly from the question’s components. The question “In adults with type 2 diabetes, does dietary counselling, compared with no intervention, reduce HbA1c levels over 12 months?” generates specific inclusion criteria (adults, type 2 diabetes diagnosis, dietary counselling intervention, comparison against no dietary intervention, HbA1c outcome measured at 12 months). Each PICO element becomes a search term and an eligibility criterion. PRISMA reporting guidelines, the standard for systematic review documentation, require the question to be formulated using a recognised structured format before the search is conducted — not derived from what the search returns.
Every type of academic paper requires a research question — but the question’s scope, methodology implication, and novelty requirement vary by paper type. An undergraduate essay question can be answerable through course readings and one or two additional sources. A postgraduate literature review question must address a defined gap in a defined literature base. A dissertation question must be answerable within one researcher’s resources. A systematic review question must be formulated before the search begins and drive every inclusion/exclusion decision. A doctoral thesis question must make an original contribution to knowledge in its field. The same formation principles apply — specificity, answerability, significance — but the scale and novelty requirements escalate at each level. See our academic writing services for support at every level and paper type.
Continue your research skills development: literature reviews · dissertation writing · research papers · critical analysis · citation and referencing · data analysis assignments · statistical analysis · capstone projects · challenging research topics · annotated bibliographies · research proposals · evidence-based practice papers