Research Question Builder

Purpose

Every study inherits its quality from its question. A vague question produces a protocol that cannot be written, a sample size that cannot be justified, and results that cannot be interpreted. This skill converts a broad topic into a structured, answerable question by forcing each element into the open: who is studied, what is done or observed, what it is compared with, what is measured, how, and over what period. It is a planning and education tool for researchers, students, and thesis writers.

Question archetypes and their frameworks

Different questions have different grammars, and the skill matches the framework to the archetype rather than forcing one template onto everything.

• Effectiveness questions (does treatment A beat treatment B?) use PICO or PICOT: population, intervention, comparator, outcome, and time. Randomized trials and systematic reviews of them live here.
• Etiology and harm questions (does exposure X raise the risk of outcome Y?) use the same slots with exposure in place of intervention, and typically name observational designs: cohort, case-control, or cross-sectional.
• Qualitative questions (what is the experience or meaning of X?) fit SPIDER better: sample, phenomenon of interest, design, evaluation, research type. Forcing a qualitative inquiry into PICO strips out what makes it answerable.
• Diagnostic accuracy questions name the index test, the reference standard, and the concerned population, because those three, not an intervention, define the study.

The one-element rule: a question element that cannot be stated specifically (a population with no age or condition boundary, an outcome with no measurement) is listed in missing_elements rather than papered over with generic words.

From structured question to study question

Structure is necessary but not sufficient. The FINER criteria are the second gate: is the study Feasible with the available population, funding, and skills; Interesting to the field; Novel relative to what is known (or a genuine replication where replication is valued); Ethical to conduct; and Relevant to practice, policy, or theory? A well-structured question that fails feasibility is a wish, not a study. The skill's output supports FINER review by making the question concrete enough to price: once the population and timeframe are explicit, the numbers needed for a sample size become arguable.

Two questions hide inside most topics: the primary question the study answers and the secondary questions it will be powered or not powered to address. Writing both down early prevents the common protocol defect of secondary analyses that are described as if they were primary findings.

Common failure modes

• Question restates the topic: "studying diabetes and exercise" is a topic, not a question with a comparator and outcome.
• Comparator omission: without a comparator, an effect cannot be defined, and the analysis section will inherit the gap.
• Outcome without measurement: "improve health" names no instrument, timepoint, or unit, so no analysis can be specified.
• Design mismatch: a prevalence question answered with a cohort design, or an accuracy question run as a trial.
• Missing_elements ignored: the skill flags gaps; treating the flags as decoration defers the same problems to the protocol stage, where they cost more.

Integration with other skills

The PICO Extractor runs this pipeline in reverse on existing abstracts, which makes the two skills useful for teaching question structure. The Inclusion/Exclusion Criteria Generator operationalizes the question into screening rules, the PubMed Search Builder turns it into database searches, and the Statistical Analysis Plan Template quantifies the primary question into endpoints and models. The Research Design and Ethics category gates every downstream step.