ThezisHow to narrow a research question: before-and-after examples
Narrow a research question by choosing one problem you can investigate with the evidence, method, and time available. Adding a country and a date can help, but a more specific sentence is not automatically a more answerable question. Start by diagnosing what makes your current question too large, then choose the kind of answer your project can support.
What makes your question too broad?
| Symptom | Likely problem | First repair |
|---|---|---|
| The outline needs several unrelated literatures. | You are investigating more than one phenomenon. | Choose the relationship or tension that matters most. |
| You ask what causes an outcome but only have interviews about experiences. | The question promises more than the design can establish. | Change the question or secure a suitable design and data. |
| The question is specific, but no usable material is accessible. | The project is too narrow or relies on unavailable evidence. | Relax one boundary or choose an accessible body of material. |
| A short factual lookup answers it. | There may be too little analysis for this assignment. | Ask about an explanation, comparison, interpretation, or unresolved relationship. |
George Mason’s writing guidance emphasizes clarity, focus, and a question that needs analysis. Treat those as checks against the assignment, rather than a rule that every discipline must use the same wording. A tightly specified experimental question may legitimately be phrased differently from a question for a literary essay. George Mason: How to write a research question
Case 1: one interest, two different projects
How does remote work affect employees?
Remote work includes many arrangements, employees includes many populations, and affect could refer to pay, health, productivity, relationships, or career progression.
Imagine a student who can interview new hires in one organization but cannot access performance records. A workable candidate is: ‘How do new hires in a hybrid team describe getting informal help during their first three months?’ This chooses a specific experience, a group, and a period. It also asks for the kind of account interviews can provide. Access, recruitment, and any required ethics approval still need confirmation.
A literature-based assignment needs a different question: ‘How do empirical studies explain access to informal help among employees working remotely?’ The student would then define eligible settings and search boundaries based on the assignment and preliminary reading. That project examines published explanations; it cannot claim to report what happened inside the student’s own organization.
Both questions came from the same interest. The available evidence changed the question’s verb and object. That is a more consequential decision than trimming a few words from the original sentence.
Case 2: narrow a literary question without making it trivial
Before: How do Victorian novels represent education? Candidate: How does the Lowood sequence in Jane Eyre connect institutional discipline with Jane’s judgments about authority?
The candidate chooses one text, a bounded sequence, and an interpretive relationship. It still requires an argument supported by close reading.
Before committing, mark the passages you would analyze and locate scholarship on the relevant concepts. You might discover that the Lowood sequence alone cannot support the argument you want to make. You could include a carefully chosen later scene to test whether the pattern persists. That expansion would have a purpose, rather than returning to a survey of the whole novel.
Do not decide the answer in the question. ‘How does the novel prove that institutions are always oppressive?’ makes a broad conclusion before the evidence has been examined. A research question should leave room for details that complicate your first reading.
Case 3: a narrow question with the wrong evidence
Did installing study booths increase final-year students’ grades at University X during the spring term?
The sentence supplies a place, group, intervention, outcome, and time. It still requires a design capable of supporting the proposed causal inference.
If you only have a one-time satisfaction survey, adding more detail will not solve the evidence problem. You could instead ask how respondents describe using the booths and which barriers they report. Or you could work with your supervisor on an appropriate evaluation design and confirm access to the necessary records. These are different projects, not interchangeable wordings.
When a project is already under way, a question change can affect the literature review, methods, and approvals. Agree the change before collecting new data. For a literature-only paper, check that the revised question still fits the assigned genre; a class review is not automatically a systematic review because you have made its scope precise.
Run a small scope trial before committing
Take the candidate question through one focused reading session. Search with a few combinations of its central concepts, inspect the most relevant records, and read enough of several sources to see how they would contribute. This is a feasibility check, not a complete literature review and not a universal minimum number of sources.
- Write two or three plausible directions an answer could take. If you can imagine only a fact to report, check whether the assignment expects more analysis.
- Sketch the sections needed to answer the question. If several sections answer different questions, choose which project you are actually doing.
- Name the evidence for each section and confirm that you can access it. An attractive outline does not make unavailable data usable.
- Write one sentence stating what the project will leave outside its scope. Explain the reason, such as relevance, design, access, or time.
If results are sparse, test synonyms and adjacent concepts before assuming there is no research. If the question yields too many relevant subtopics, choose a more consequential boundary. Your goal is enough material to analyze within the assignment, not the smallest possible search result count. Check whether you have a research gap
My current question is … I can answer it using … The main boundary is … because … My preliminary reading suggests these competing explanations … The decision I need help with is …
Fill this in with actual sources and constraints. A specific decision is easier to discuss than ‘Is my topic good?’
Once the scope is agreed, put its reading, drafting, and feedback milestones into your thesis plan. Keep the question visible while writing so that new material earns its place. Plan your bachelor thesis