How to Write a Research Question

A good research question is a single, clear sentence that names the population, the exposure or intervention, the comparator (if any) and the outcome, and that your study can answer with the data you will collect. It determines the design, the sample and the analysis, so time spent here saves effort later.

Qualities of a good question

• Focused: one primary outcome.
• Answerable: measurable with available methods.
• Relevant: matters to patients, clinicians or policy.
• Ethical and feasible.
• Not already answered.

Templates by type

• Intervention: "In [population], does [intervention] compared with [comparator] change [outcome] over [time]?"
• Prognosis/risk (observational): "Among [population], is [exposure] associated with [outcome]?"
• Frequency: "What proportion of [population] in [setting] has [condition]?"
• Diagnosis: "In [population], how accurate is [test] compared with [reference standard] for [condition]?"
• Qualitative: "How do [participants] experience [phenomenon] in [setting]?"

Illustrative example

Invented for teaching. Weak: "Does exercise help the heart?" Better: "In adults aged 50–70 with hypertension (P), does a 12-week supervised aerobic programme (I) compared with usual care (C) lower resting systolic blood pressure (O)?" The revision names who, what, compared with what, and how the result is measured.

Revising your question with a mentor

Share the one-sentence question and ask three things: Is it clinically meaningful? Is it answerable with the data we can access? What is the single primary outcome? Mentors often spot that a question mixes two studies in one. Expect to revise the question two or three times; a question that survives these revisions is much more likely to lead to a finished project. MRN's PICO worksheet can guide the revisions.

Common pitfalls

• Two or more questions in one sentence.
• Vague words: "improve", "affect", "outcomes".
• A question that is really a hypothesis or a title.

Key takeaways

• One question, one primary outcome, all elements named.
• Use the template that matches your study type.
• Test the draft: could a colleague tell what data you would collect?

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.