03 · What You Need to Know
A Main Question and Subquestions Create a Hierarchy
The defining feature of this structure is hierarchy. The main question states the overarching inquiry. Subquestions divide that inquiry into more specific areas that need investigation.
Qualitative methods literature makes this distinction particularly explicit. Busetto and colleagues describe qualitative research questions as generally involving a central question and related subquestions: the central question addresses the primary phenomenon, while subquestions examine more specific areas of focus. Barroga and Matanguihan similarly describe qualitative research as commonly using a central question with associated subquestions rather than relying on hypotheses.
The structure is not exclusive to qualitative research. Larger quantitative, mixed-methods, case-study, evaluation, and dissertation projects can also benefit from hierarchical questions. The essential logic remains the same: answering the smaller questions should help answer the larger one.
What is a main research question?
The main research question, sometimes called the central or overarching question, expresses the principal inquiry that organizes the study.
For example:
“How do first-year university students navigate institutional expectations concerning generative AI in academic writing?”
This question identifies the phenomenon broadly enough to accommodate several dimensions of students' experiences while still establishing a recognizable focus.
In qualitative inquiry especially, a central question may intentionally remain broad and open-ended so that participants' perspectives are not prematurely restricted. Methodological guidance describes the central question as addressing the primary phenomenon, with subquestions providing more detailed areas of inquiry.
What is a research subquestion?
A research subquestion asks about a particular part of the main question.
For the example above, possible subquestions might be:
“How do students interpret institutional and instructor guidance concerning acceptable generative AI use?”
“What considerations shape students' decisions about when to use generative AI?”
“How do students respond when expectations differ across courses?”
These questions are narrower than the main question but remain inside its conceptual boundaries. Together, their answers help build an answer to the overarching inquiry.
Creswell and Poth describe subquestions in qualitative research as questions that further specify a central question into areas for inquiry. Their function is essentially to subdivide the central question into constituent parts.
Main or central question
States the overarching inquiry the study ultimately seeks to answer.
Subquestion
Investigates a defined component of that overarching inquiry whose answer contributes to answering the main question.
The simplest test is whether answering the subquestions answers the main question
Imagine completing your analysis and obtaining convincing answers to every subquestion. Could you then construct a substantive answer to the main question?
If yes, the hierarchy may be working.
If not, something is wrong. Perhaps an important dimension of the main question has no corresponding subquestion. Perhaps some subquestions wander beyond the main inquiry. Or perhaps the main question is so broad that no realistic set of subquestions can adequately answer it.
Academic guidance on central questions and subquestions makes this relationship explicit: subquestions should address subtopics of the central theme, and answering them collectively should allow the central question to be answered.
Subquestions can turn a complex question into manageable analytical work
A good main question may be conceptually coherent but still difficult to investigate as one undivided task.
Consider:
“How has a university's transition to hybrid learning changed students' experiences of academic participation?”
The researcher might need to investigate participation in classroom interaction, access to learning activities, peer collaboration, and experiences of belonging. Subquestions can make those areas explicit without pretending they are independent studies.
This is one practical advantage of hierarchy. Subquestions can transform an overarching intellectual problem into more manageable analytical tasks while preserving the connection among them.
This should not be confused with simply adding more research questions until the project becomes unmanageable. Subquestions should organize complexity rather than manufacture it.
Subquestions are especially useful when the main question has several necessary dimensions
A main question may require investigation of several aspects before it can be answered adequately.
For example:
Main question: “How is a university-wide generative AI policy implemented across academic departments?”
Possible subquestions could examine:
- how departments interpret the institutional policy;
- how those interpretations are translated into local practices;
- where implementation differs among departments; and
- what conditions participants identify as shaping those differences.
The subquestions do not merely create additional topics. They identify parts of the implementation process that collectively illuminate the central question.
Subquestions can separate perspectives within one phenomenon
Some research problems require evidence from different participant groups.
Suppose the main question is:
“How is a new university assessment policy experienced during implementation?”
Subquestions might examine students' experiences, instructors' experiences, and administrators' experiences.
This can be coherent when those perspectives are intentionally brought together to understand the same implementation phenomenon.
However, each additional population increases the study's methodological burden. If the student, instructor, and administrator questions develop separate theoretical rationales, datasets, analyses, and conclusions with little integration, they may be better understood as distinct research questions or even separate studies.
Subquestions can reflect stages of a process
Process-oriented studies may use subquestions to examine different stages of an unfolding phenomenon.
For example:
Main question: “How do universities develop and implement policies for generative AI in assessment?”
Subquestions might ask how the issue first enters institutional decision-making, how policy options are developed, how stakeholders participate, how the policy is communicated, and how implementation problems are addressed.
The sequence matters because the subquestions collectively reconstruct the process.
Creswell's discussion of qualitative subquestions similarly distinguishes questions about substantive issues from procedural questions that can help investigate how a process unfolds, who participates, what events occur, and what outcomes emerge.
Subquestions should narrow the main question without changing the study
Suppose your main question concerns how students experience generative AI policies.
A subquestion about how students interpret policy language fits naturally.
A subquestion asking whether generative AI causally improves examination performance probably does not. That question requires a different form of evidence and answers a different research problem.
The fact that both questions mention generative AI does not make one subordinate to the other.
Watch Out
Do not use a broad main question as an umbrella for unrelated studies. A subquestion should investigate part of the main inquiry, not merely another interesting issue involving the same topic.
The main question should not simply repeat the subquestions in one sentence
Another common structure is:
Main question: “What are students' AI-use frequency, attitudes toward AI, writing self-efficacy, academic performance, and perceptions of institutional AI policies?”
Then five subquestions ask separately about those five items.
This is technically hierarchical, but the main question adds little beyond bundling the subquestions together.
A useful main question should express the larger intellectual inquiry that connects the parts. If the components have no meaningful relationship beyond appearing in the same questionnaire, there may be no genuine overarching question.
Subquestions are not the same as interview questions
This distinction is particularly important in qualitative research.
A research subquestion expresses something the study seeks to answer. An interview question is something you actually ask a participant to generate data relevant to that inquiry.
For example:
Research subquestion: “How do students interpret institutional expectations concerning acceptable generative AI use?”
Possible interview questions might include:
“What guidance about generative AI have you received from the university?”
“Can you describe a situation in which you were unsure whether AI use was allowed?”
“How did you decide what to do?”
The interview questions are more concrete and conversational. Several interview questions may contribute evidence to one research subquestion.
Creswell's methodological guidance notes that qualitative subquestions can inform the core questions used during interviews, observations, or document analysis. That relationship does not make the two levels interchangeable.
Subquestions are not automatically secondary research questions
The terms “subquestion” and “secondary research question” are sometimes used loosely as though they mean the same thing. They can serve different purposes.
| Structure |
Relationship |
Typical use |
| Main question + subquestions |
Subquestions divide the overarching question into components needed to answer it |
Common in qualitative, case-study, dissertation, and other complex inquiries |
| Primary + secondary questions |
The primary question has greater inferential or design priority; secondary questions address additional uncertainties |
Common in confirmatory quantitative, intervention, and clinical research |
| Several co-equal research questions |
Each question addresses a distinct but connected aspect of the research problem without necessarily being subordinate to another |
Possible across many designs when the questions genuinely have comparable status |
In practice, terminology varies across disciplines and institutions. The conceptual distinction is more important than the label.
The previous guide on whether a study should have one primary research question addresses the primary-secondary structure more directly.
Qualitative research commonly uses central questions and subquestions
The central-question structure is particularly established in qualitative methodology. Qualitative research questions are often open-ended and may evolve as understanding develops. A central question provides direction without prematurely reducing the phenomenon to a narrow set of variables, while subquestions focus attention on particular dimensions.
This makes the structure useful when a phenomenon is complex but still conceptually unified.
For example:
Central question: “How do doctoral students experience the transition from coursework to independent research?”
Subquestions might examine changes in researcher identity, relationships with supervisors, experiences of uncertainty, and strategies for developing independence.
The researcher is not conducting four unrelated studies. The subquestions provide lenses through which the central transition can be understood.
Do qualitative studies need exactly five to seven subquestions?
No.
Some qualitative methods guidance recommends a small number of subquestions and gives five to seven as a practical guideline. Busetto and colleagues report this recommendation, and Creswell and Poth likewise suggest a small number, commonly five to seven, to refine a central question.
That is methodological guidance, not a universal numerical requirement.
A study may need two strong subquestions. Another may need five. Adding weak questions merely to reach a recommended count defeats the purpose of the structure.
The more important test is whether the subquestions collectively cover the aspects of the main question that the study actually needs to investigate without expanding beyond what can be answered well.
Quantitative research may not need subquestions at all
A focused quantitative study may be better represented by one primary research question and one or more secondary questions or hypotheses.
Suppose a randomized study asks:
“Does retrieval-practice instruction improve delayed recall compared with rereading?”
Breaking that question into several subquestions may add no value. The design may be clearer when organized around one primary question, its corresponding hypothesis, and any prespecified secondary outcomes.
Research-question structure should therefore follow the methodological logic rather than a thesis template copied from another discipline.
Mixed-methods research may need a different hierarchy
Mixed-methods studies often contain distinct quantitative and qualitative questions because the two strands seek different kinds of evidence.
For example:
Quantitative question: “How frequently do undergraduate students use generative AI for academic writing?”
Qualitative question: “How do students explain their decisions about when to use generative AI?”
An overarching mixed-methods question might then ask how students' explanations help interpret the quantitative patterns.
In that design, calling the qualitative question a “subquestion” of the quantitative question may misrepresent their relationship. They may instead be complementary strands whose findings are integrated.
This is why questions requiring different methods should be organized according to the logic of the overall design rather than forced into an artificial hierarchy.
Subquestions can expose an overly broad main question
Trying to formulate subquestions is itself a useful diagnostic exercise.
Suppose your main question is:
“How does technology affect education?”
You begin writing subquestions and quickly produce questions about artificial intelligence, learning management systems, virtual reality, teacher workload, student achievement, academic integrity, educational access, institutional finance, and policy.
The problem is not that you need eight subquestions. The main question is probably too broad.
A coherent central question should generate a bounded family of subquestions. If almost any issue in the field could become a subquestion, revisit whether the main question has sufficient scope.
Subquestions can also reveal an overly narrow main question
The reverse can happen.
Main question: “What percentage of first-year students used generative AI at least once during the previous seven days?”
There may be little reason to create several subquestions. The main question already asks for one tightly defined estimate.
If you find yourself inventing subquestions about attitudes, motivations, academic performance, and policy merely because a proposal template expects them, you are no longer decomposing the main question. You are expanding the study.
A question can be appropriately narrow without requiring further subdivision.
Every subquestion should have an evidentiary path
A well-worded subquestion still fails if the study cannot answer it.
For each subquestion, identify:
- what evidence would answer it;
- which participants, cases, documents, or measurements provide that evidence;
- how the evidence will be generated or obtained; and
- how it will be analyzed.
If the main question concerns student experiences but one subquestion requires administrative financial records that the study will never collect, the hierarchy is methodologically incomplete.
Subquestions are therefore not merely an outline for the literature review. They make claims about what the empirical study intends to answer.
The subquestions should be distinct without becoming independent
Good subquestions occupy a useful middle ground.
If two subquestions ask essentially the same thing, consolidate them.
If a subquestion can be removed without changing your ability to answer the main question, ask whether it is genuinely necessary.
If a subquestion requires an entirely separate research problem, theoretical framework, population, and methodology, consider whether it belongs in the study at all.
A strong hierarchy therefore has both differentiation and dependency: each subquestion contributes something distinct, yet each remains dependent on the main inquiry for its reason to exist.
The final answer should return to the main question
A hierarchical structure creates an important obligation during interpretation.
The study should not finish with separate answers to SQ1, SQ2, SQ3, and SQ4 and leave the reader to assemble them. The researcher needs to integrate those answers and return explicitly to the main question.
If the subquestions are the rooms, to borrow one university's useful analogy, the main question is the house they collectively constitute.
The intellectual contribution lies partly in explaining what the combined findings mean for the overarching inquiry.
