Research Question: How to Write One That Works

A research question is the precise, answerable question your study exists to answer. It names the population or setting, the factor or experience you will explore, the outcome, and often the time frame. That wording drives everything downstream: the databases you search, your inclusion criteria, your data collection plan, and the information you extract. This article, like our other guides, covers question types, from clinical comparisons to open questions in qualitative research, and shows how to test a draft research question before you commit to a method.

Key takeaways

• A research question is the single sentence that names what you want to find out, and it sets the boundaries for your search, your sample, and your analysis.
• Choose the question type before the method: descriptive, comparative, relational, evaluative, and exploratory questions each commit you to a different type of study.
• Structured frameworks help you test a draft rather than write it. PICO suits comparative clinical questions, SPIDER suits qualitative work, and FINER checks feasibility and relevance.
• Most weak questions fail for one of three reasons: they are too broad to answer, they smuggle in the answer you expect, or nobody has produced the data needed to answer them.
• Score your draft against a short checklist before you write a protocol, because a vague question costs the most when it is discovered during screening or data extraction.

What Is a Research Question and What Does It Do?

The question at the center of a study is a single, answerable sentence that names exactly what the work will find out. It narrows a broad topic into something you can investigate, and it does four jobs at once. It defines scope, so you know which literature belongs. It implies a method, because a question about prevalence and one about causation need different designs. It sets the unit of analysis, which decides what information you extract from each study. And it fixes the outcome, turning a vague interest into a testable claim.

Strong questions share a few characteristics. They are focused, they can be answered with existing data or resources, and they are tied to a real problem rather than general curiosity. A hypothesis is different. The question asks what is true, while the hypothesis predicts what you expect to find and can be wrong.

A pair of PhD examples shows the range. "How do first generation undergraduates interpret academic feedback during their first semester?" is exploratory, suits interviews, and brings out issues such as confidence and belonging. "Does structured feedback training reduce first year attrition?" commits you to comparison and measurement. Both work at doctoral scale, but each needs different evidence.

Types of Research Questions Used in Qualitative Studies

Most guides list five types of research questions: descriptive, comparative, relational, evaluative, and interpretive. Qualitative studies rely mostly on the interpretive and descriptive types. These ask how people understand an experience or what a process looks like from the inside, so they call for interviews, focus groups, or observation. Comparative, relational, and evaluative questions usually need quantitative data, controlled designs, or enough cases for statistics.

Whatever its type, a good research question has the same characteristics. It is focused, answerable with the methods you have, grounded in existing literature, ethically feasible, and specific about population and setting. Compare "What do patients think of remote consultations?" with "How do older adults with type 2 diabetes describe trust in remote consultations?" The second question names the group and the experience, so it points straight to a sampling plan and an interview guide. Let the literature guide your choice among types of research questions. If several qualitative studies already exist, a synthesis may add more than another small study, as the systematic review methodology guide explains.

Common Mistakes to Avoid When Framing Your Question

Most weak questions share a few faults. The two-in-one question joins separate concerns with "and" and then needs two designs to resolve. The loaded question assumes its own conclusion and biases the search. The unanswerable question has no dataset and no realistic way to collect one, a problem that often surfaces late in a project.

• Name the outcome. "Better outcomes" lets every reviewer read the question differently; PICO forces each element into the open.
• Match the wording to the aim. A descriptive question asks what exists, while a question about impact asks whether one factor changes another.
• Leave your preferred intervention out of the question.
• Define the population tightly enough for one study to recruit it.

Test the wording against these faults before drafting a protocol.

How to Narrow a Broad Topic Into a Focused Inquiry

Narrowing works best as a series of deliberate cuts rather than one leap. Start from the broad problem, then reduce it along four axes in order: population, setting, outcome, and time. "Burnout in healthcare" becomes "burnout in emergency department nurses," then "burnout in emergency department nurses working rotating night shifts," then "changes in burnout scores over twelve months among those nurses." Each cut makes the work more feasible and the results more interpretable.

Read before you narrow. A scan of recent reviews shows which slices are already saturated and where an honest gap sits, and it tells you whether measurement tools exist for the outcome you want to study. That reading is also where a researcher finds the validated instrument that makes a hypothesis testable rather than rhetorical.

Templates speed this up. Worked examples of PICO questions give you a ready structure for comparative clinical work, and reviewing several before you write a research question shows how tight the wording needs to be. Where you plan to investigate experiences instead of effects, keep the population narrow and let the outcome stay open. Good research usually comes from a modest question answered well, not a large one answered loosely.

Using FINER and PICO to Test Your Draft Question

Apply FINER concretely. Feasibility means counting eligible cases at your site and checking whether recruitment math works. Novelty means running a quick search rather than trusting memory. Ethics means asking whether the comparator withholds effective care. A question that clears all five is usually strong research territory; one that fails feasibility should be rescoped, not abandoned.

For content structure, work through several examples of PICO questions to see how each element is named. PICO also indicates the type of study you need and the analysis that follows, since a stated comparator implies a control group and an effect estimate. If your draft has no comparator, you are likely asking a descriptive or interpretive question, and forcing PICO onto it will distort the aim. Our PICO templates cover the mechanics; a hypothesis follows once the four elements are fixed.

How Specific and Focused Should Your Question Be?

A research question should be specific and focused enough that two competent reviewers would sort the same set of papers the same way, and no more specific than your data can support. That is the practical test. If a colleague reads your question and cannot predict which studies you would include, the wording is too loose. If the question is so narrow that fewer than a handful of eligible cases exist, you have overspecified it and the analysis will be underpowered.

Balance matters. Over-narrowing is a real failure mode: restricting to one hospital, one year, and one age band can leave you with an answer that holds nowhere else. The purpose of specificity is replicable judgment, not minimal scope. A useful middle setting names the population and outcome precisely while leaving the setting broad enough to recruit.

Practically, write the question, then write the inclusion criteria straight from it. Any criterion you cannot trace to a word in the question shows a gap you should close. Examples of PICOT questions for nursing research demonstrate this well, because the added time element removes ambiguity about the follow up window. Comparing a PICO and SPIDER research question on the same topic also shows how much precision each framework expects. Strong research questions leave the hypothesis obvious and the analysis already half planned before you ask anyone for data.

Weak Versus Strong Question Examples Side by Side

Comparing drafts side by side is the fastest way to see what precision buys you. Each weak version below fails on a specific point, and the revision fixes it without changing the underlying research topic.

• Weak: "Does social media affect teenagers?" Strong: "Is daily time on image based social media associated with body dissatisfaction scores among US girls aged 13 to 17?" The revision names population, exposure, and outcome.
• Weak: "Is telehealth good?" Strong: "Do telehealth follow ups reduce 30 day readmissions after heart failure discharge compared with clinic visits?" The revision adds a comparator so the investigation can produce an effect estimate.
• Weak: "What are the problems with nurse staffing?" Strong: "How do ward nurses describe rationing care during understaffed night shifts?" The revision turns a vague issue into an answerable interpretive question.

Apply the same discipline to a thesis. Compare PICO and SPIDER versions of your question, or review examples of PICOT questions for nursing research, then ask which wording a reader could act on without further explanation. Good research questions survive that test; weak ones need a paragraph of clarification before anyone can answer them.

Self-Assessment Checklist for Scoring Your Question

Score your draft on seven points, one mark each, before you write anything else. A draft scoring five or below needs another pass.

• Clarity: a reader outside your field can state what you will find out without asking a follow up question.
• Single focus: the wording contains one specific question, not two joined by "and."
• Researchable: the data either exists or can be collected within your timeline and budget.
• Measurable outcome: you can name the instrument, scale, or coding approach.
• Neutral framing: the wording does not presuppose the answer.
• Contribution: the question adds knowledge rather than repeating a settled issue.
• Traceability: every inclusion criterion maps to a word in the question.

Two extra checks pay off in review work. First, pilot screen twenty records against your criteria; disagreement above roughly one in five usually means the question, not the reviewers, is at fault. Second, read examples of PICO questions for nursing research and examples of PICOT research questions alongside your draft to confirm each element is explicit. Good research questions, whether for a thesis or a synthesis, pass all seven points before a protocol is written.

Why a Sharp Question Decides the Success of Your Project

The question determines what a project can produce, and no later effort compensates for a weak one. A sharp question tells you when to stop searching, which records to keep, and which numbers to report. A loose one leaves every decision negotiable, which is why vague projects expand: without a boundary, each new paper looks arguably relevant. Teams that fix the wording first tend to finish; teams that leave it open tend to rescope halfway through.

There is a cost curve worth knowing. Rewriting a question during planning costs hours. Rewriting it after screening means redoing the search and the eligibility log. Rewriting it after data extraction usually means restarting the project. That asymmetry is the strongest practical argument for spending a week on wording before you write a research protocol.

A researchable question also protects the write-up. Your methodology section becomes a direct account of how you answered one stated aim, your results answer that aim in the same terms, and each reference in the discussion connects to it. Exploratory projects still need this discipline; they simply state the aim as territory to map rather than an effect to estimate. Tools such as Systematicly can keep the question, criteria, and extraction fields aligned across a project, but the knowledge gain comes from the question itself.

How to Develop a Research Question Step by Step

To make a question, start broad and narrow it in deliberate steps. Each step should leave you with something you can write down and check against the literature.

1. Pick a topic and state the problem. Name the population, condition, or practice you care about and explain why the current evidence feels unsettled.
2. Run a scoping search. Search one or two databases to explore what has already been reviewed. If a recent review already answers your question, sharpen it or choose a different angle.
3. Apply a framework. Break the question into elements, such as PICO (population, intervention, comparison, outcome) for effectiveness questions or SPIDER for qualitative work.
4. Draft and test the wording. Write one sentence. Check that it is specific enough to judge study eligibility and broad enough to return studies.

The finished research question then drives the rest of the research process. Each element becomes a search concept, an inclusion criterion, and a data extraction field, so a vague question produces a vague search. Record the final wording in your review protocol before screening begins. That record keeps the research process transparent when you report under PRISMA.

Choosing a Broad Subject Area as Your Starting Point

Most researchers start with a subject they already care about, such as sleep in shift workers or remote learning in rural schools. A broad area is not yet a question, but it tells you where to look. Before narrowing, spend a few hours scanning recent reviews and trial registries to see what has been studied, what conflicts, and what is missing.

Test the area against three practical limits:

• Relevance: Would an answer change practice, policy, or the impact of a known intervention?
• Feasibility: Can you realistically handle the data collection or screening load within your time and budget?
• Method fit: Does the topic lean toward measuring effects, or toward exploring experiences, which shapes the kinds of questions qualitative research can answer?

Our systematic review guides show how this early scan feeds a protocol.

Sample Questions That Show the Concepts in Practice

A well-built research question states who, what, and which outcome clearly enough to screen records against it. Here are three examples:

• PICO: In adults with type 2 diabetes, does a structured exercise program, compared with usual care, lower HbA1c at 12 months?
• PEO: Among night shift nurses, is long-term shift work associated with sleep disorders?
• SPIDER: How do family caregivers of people with dementia describe respite care in interview-based qualitative research?

Each version gives researchers eligibility criteria and search terms they can use right away. The qualitative example replaces outcomes with experiences because its context is lived experience, not effect size. A vague version such as "Does exercise help diabetes?" leaves out the population, comparator, and time frame, so screening decisions become inconsistent. Our systematic review guides show how to turn questions like these into a protocol and a study selection plan.

Applying the FINERMAPS Criteria to Test Your Question

FINERMAPS extends the familiar FINER checklist into a stress test for a question before you register a protocol. Score each criterion yes, no, or unclear:

• Feasible, Manageable: you have enough studies, time, and reviewers to screen them.
• Interesting, Novel, Relevant: no recent review already answers it, and the answer would matter to clinicians or policymakers.
• Ethical, Appropriate: the review raises no ethical concerns, and the design suits the question type.
• Potential value, Publishability, Systematic: the findings fill a real gap and can follow a reproducible method.

An "unclear" score usually means you need more information from a scoping search. Our systematic review guides show how to narrow a question that fails these checks.

What Makes a Study Question Strong and Workable?

A strong research question names exactly what you will examine and what evidence would answer it. It should pass four tests. It is specific, defining the population, intervention or exposure, comparison, and outcome, so two reviewers would judge the same study the same way. It is answerable, which a quick scoping search can confirm. It is feasible within your time, budget, and skills. And it matters to the people who will use the result.

Workability comes down to scope. A question that is too broad returns thousands of records you cannot screen, while one that is too narrow may surface only a single study. Draft eligibility criteria directly from each element and check that every study you include could, in principle, change the answer. Finally, match each study design you accept to the question type.

Practical Tips for Writing Your Question Clearly

A clear research question says exactly what you will study and how you will know when you have an answer. Use these habits when writing a good research question:

• Start with a framework. Map the parts with PICO (population, intervention, comparison, outcome) for effectiveness questions. Use SPIDER or PEO for questions in qualitative research, which focus on experiences rather than measured effects.
• Keep one main aim. Words like "does," "how," or "what" should lead to a single answerable aim, not three aims bundled together.
• Define every term. Replace vague words like "young" or "improve" with age ranges, settings, and named outcomes.
• Check feasibility. Run a quick scoping search to confirm enough information exists without pulling in thousands of off-topic records.

Writing down the question before screening starts matters as much as wording it well. Put the final version in your review protocol and date it. If you change it later, log the change and the reason. That record shows reviewers that your eligibility criteria came from the question and were not adjusted to fit the results.

Frequently asked questions

What are examples of research questions?

Most strong examples follow one pattern: they name a population, a factor, and an outcome. "Do standing desks reduce reported lower back pain among office workers over six months?" is a comparative question. "How do community pharmacists describe deprescribing conversations with older adults?" is interpretive. "What proportion of high school teachers use generative AI for lesson planning?" is descriptive.

What are the 5 good research questions?

The five characteristics of a good research question work as a checklist. A good question is clear, so two readers interpret it the same way. It is focused on one issue and feasible with your time, data, and access. You can answer it with evidence you can measure or observe. It is also relevant, adding something the existing literature has not already settled.

What is a research question?

A question is the specific, answerable question a study sets out to resolve. It matters to the research process because it determines the design, the data you collect, the search strategy for a review, and the analysis. Without one, your method choices drift.

What is a strong research question?

A strong research question is narrow enough to answer and important enough to matter. To write one, start with a broad topic, read recent studies to find a gap, and state the problem. Then draft the question with a framework like PICO. Test it by asking whether two reviewers would apply it to the same studies and reach the same decisions.

What are 5 good research questions?

Here are five good questions on one topic, night shift nursing: What is the prevalence of sleep deprivation among night nurses? Do 12 hour shifts raise medication error rates compared with 8 hour shifts? How do night nurses describe managing fatigue? Does a scheduled nap protocol reduce errors over three months? How do fatigue scores change across a six month rotation?

What is a good research example?

A good research example traces the full chain. Topic: sleep in nursing. Problem: night shift nurses report fatigue, but its link to errors is unclear. Question: do nurses on 12 hour night shifts make more medication errors than nurses on 8 hour shifts? Design: a cohort study, or a meta-analysis if comparable studies already exist.

Can you give me an example of a research problem?

Suppose a rural health system offers remote monitoring to heart failure patients, yet readmissions have not dropped and staff do not know why. That is the research problem: a gap between what was expected and what happened. A research question then narrows it. For example: does remote monitoring adherence predict 30 day readmission among rural heart failure patients?

What is in a research question?

A complete question contains a defined population, an exposure, intervention, or experience, a measurable outcome, and usually a setting or time frame. Comparative questions also name a comparison group. Frameworks such as PICO help you check that nothing is missing.

What are 5 good research topics?

Five topics with room for new work are digital mental health support for adolescents, antimicrobial stewardship in primary care, teacher retention in rural schools, food insecurity among college students, and reporting quality in published systematic reviews. Each one still needs narrowing into a question.

What are 10 examples of research titles?

Ten example titles: Remote Monitoring and Heart Failure Readmissions; PICO Framing in Nursing Reviews; Sleep and Academic Performance; Peer Mentoring and First Year Retention; Telehealth Uptake in Rural Clinics; Nurse Staffing and Missed Care; Screening Reliability in Evidence Synthesis; Deprescribing in Community Pharmacy; Standing Desks and Back Pain; and Generative AI in Lesson Planning.

What are 10 broad research topic areas?

Ten broad areas are public health, education policy, clinical nursing, mental health, environmental science, labor economics, nutrition, digital technology use, health equity, and research methods. Where an area already has a large published literature, a systematic review can answer a question by pooling existing studies instead of collecting new data.

What are the top 10 research topics for high school students?

Ten topics where the data is easy to get are screen time and sleep, part time work and grades, school start times, cafeteria food waste, local air quality, recycling behavior, exercise and mood, study technique effectiveness, social media use and test anxiety, and household water use.