How to Write a Strong Research Question

What Is a Research Question?

A research question is a clearly formulated question that states what a study seeks to investigate, understand, describe, compare, explain or evaluate.

It is more specific than a research topic and more operational than a general research problem.

Consider this progression:

The broad topic identifies an area of interest. The research problem explains why investigation may be needed. The research question identifies the inquiry the study will actually address.

A well-formulated research question therefore provides direction for later decisions about research design, evidence, sampling or analytical coverage, data collection or acquisition, and analysis.

Methodological caution: A research question does not need to contain every detail of the eventual methodology. Its purpose is to define the central inquiry clearly, not to compress the entire research proposal into one sentence.

What Makes a Strong Research Question?

There is no single wording formula that works for every social-science study. However, strong research questions usually share several characteristics.

1. Clear

The reader should understand what the study intends to investigate.

Avoid vague expressions such as:

• various factors;
• different issues;
• overall impact;
• many challenges; or
• several aspects.

These expressions may conceal rather than clarify the intended inquiry.

Too vague: What factors affect employee performance?

Clearer: What is the relationship between perceived supervisor support and employee performance among administrative employees in selected public organisations?

2. Focused

A question should be narrow enough to investigate meaningfully within the scope of the study.

Too broad: How does education affect development?

This could involve educational access, attainment, quality, expenditure, employment, productivity, income, inequality, social mobility and many other dimensions.

A more focused version might be:

More focused: How is public expenditure on secondary education associated with secondary-school completion rates across selected lower-middle-income countries?

Narrowing the question does not mean adding arbitrary boundaries. Population, location, time period or other limits should be included when they genuinely define the inquiry.

3. Researchable

A research question should be answerable using evidence that can reasonably be obtained and analysed.

Depending on the study, that evidence may come from:

• surveys;
• interviews;
• observations;
• administrative records;
• policy documents;
• existing datasets;
• organisational records;
• archival materials; or
• a combination of sources.

A question may be interesting but still unsuitable for a particular project if there is no realistic way to obtain evidence capable of addressing it.

4. Feasible

A good question must also fit practical constraints.

Consider:

• available time;
• access to participants or data;
• research skills;
• financial resources;
• available software or technical expertise;
• ethical requirements; and
• the scale of the proposed analysis.

The FINER framework—Feasible, Interesting, Novel, Ethical and Relevant—is commonly used as one way of evaluating proposed research questions.

For student research in particular, feasibility is often what separates an interesting idea from a workable project.

5. Relevant

A research question should address something worth investigating.

Relevance can arise from:

• a gap or unresolved issue in existing research;
• a theoretical question;
• inconsistent previous findings;
• an important practical problem;
• a policy concern;
• a new context;
• a replication need; or
• a need to extend existing evidence.

A study does not have to discover something completely unprecedented to be worthwhile. Replicating earlier work in another setting, testing whether an established relationship holds for a different population, or reassessing previous findings using improved data can also make useful contributions.

6. Methodologically aligned

The question must be compatible with the type of evidence and research design available.

For example, observing that two variables are statistically associated does not by itself establish that one causes the other.

Words such as effect, impact, influence, causes and leads to can imply causal claims. Their use should therefore be supported by a research design and analytical strategy capable of addressing causality.

Methodological caution: If your study can establish an association but not a causal effect, wording such as “What is the relationship between…” may be more defensible than “What is the impact of…”

How to Develop a Research Question Step by Step

Developing a research question is usually iterative. The first version is rarely the final one.

Step 1: Start with a broad research interest

Begin with an area that interests you or relates to a meaningful academic or practical problem.

Examples include:

• Management: employee motivation and leadership;
• Economics: government expenditure and economic growth;
• Public Policy: implementation of social-protection programmes;
• Education: digital learning and student engagement;
• Development Studies: access to public services and household welfare.

At this stage, you are identifying an area of inquiry—not yet writing the final research question.

Step 2: Understand the research problem

Ask why the topic deserves investigation.

A research problem may emerge because:

• something is not well understood;
• existing studies disagree;
• evidence is outdated;
• a theory has not been tested in a relevant context;
• a programme or policy produces uncertain outcomes;
• a practical problem requires systematic investigation; or
• an existing relationship requires further explanation.

Reviewing the literature is essential here because useful research questions are developed in relation to what is already known, not in isolation from existing knowledge.

For example:

Broad interest: employee motivation.

More specific problem: the organisation has introduced new supervisory practices, but it is unclear whether employees who perceive greater supervisor support also report higher motivation.

You now have a much clearer basis for a question.

Step 3: Decide what you mainly want to find out

The wording of a research question depends heavily on the purpose of the inquiry.

Ask:

What do I mainly want to find out?

You may want to:

• describe a phenomenon;
• explore experiences or perspectives;
• compare groups or situations;
• examine a relationship between variables;
• evaluate a programme, policy or practice;
• explain why something occurs; or
• investigate a causal effect, where the design permits it.

This distinction matters because different questions lead toward different methodological choices.

Step 4: Identify the central concepts or phenomenon

Specify what the question is actually about.

For quantitative studies, this may involve clearly identifying variables or measurable constructs.

For qualitative studies, the focus may instead be a phenomenon, experience, process, meaning or context.

For example:

Broad: How do employees feel about work?

More focused: How do early-career employees describe their experiences of organisational support in public-sector workplaces?

The second question identifies a central phenomenon while leaving space for participants’ experiences to emerge.

Step 5: Define the relevant population, cases or analytical units

Determine who or what the question concerns.

Depending on the study, the unit may be:

• individuals;
• households;
• employees;
• organisations;
• schools;
• municipalities;
• countries;
• policies;
• firms;
• documents; or
• observations over time.

For example:

Initial question: What is the relationship between supervisor support and employee motivation?

Refined question: What is the relationship between perceived supervisor support and employee motivation among early-career employees in selected public organisations?

The refinement makes the intended analytical population clearer.

Step 6: Consider what evidence could answer the question

Before finalising the wording, ask:

What evidence would I need to answer this question?

A study can use primary data, secondary data or both.

The important point is that the evidence should be suitable for answering the question—not selected simply because it is convenient.

Step 7: Check the scope

Now ask whether your proposed question is realistic for the project you are undertaking.

A master’s dissertation, undergraduate thesis and multi-year funded research programme may reasonably operate at very different scales.

If answering the question requires:

• inaccessible data;
• several unrelated populations;
• numerous countries;
• many different outcomes;
• multiple research designs; or
• resources beyond those available,

the question may need to be narrowed.

Step 8: Draft, evaluate and revise

Write your initial question.

Then test it against the following criteria:

• Is it clear?
• Is it focused?
• Is it researchable?
• Is it feasible?
• Is it relevant?
• Is the wording neutral?
• Can the intended research design realistically answer it?

Discuss the question with a supervisor, lecturer, colleague or subject expert where appropriate.

Research questions can also be refined after further literature review or after discovering limitations in data availability. Revision is a normal part of research planning, not evidence that the original idea failed.

Common Types of Research Questions

Research questions can serve different purposes.

Descriptive questions

These ask what exists, how much, how often, or what characteristics are present.

Example: What are the reported levels of job satisfaction among administrative employees in selected public organisations?

Exploratory questions

These seek deeper understanding of an experience, process or relatively underexplored phenomenon.

Example: How do early-career teachers experience the introduction of digital learning technologies?

Relational questions

These examine associations between concepts or variables.

Example: What is the relationship between perceived organisational support and employee commitment?

Comparative questions

These examine differences or similarities between groups, settings or periods.

Example: How do reported job-satisfaction levels differ between permanent and contractual employees?

Evaluative questions

These assess how a programme, policy, intervention or process is functioning or performing.

Example: How do local government officials assess the implementation of a newly introduced digital service-delivery system?

Explanatory or causal questions

These seek to explain why something occurs or whether one factor changes another.

Example: What effect does participation in a structured employee-training programme have on subsequent job performance?

Causal questions require particular methodological care because their conclusions depend on the research design, comparison strategy and assumptions used.

Research questions may also evolve differently in quantitative and qualitative inquiry. Quantitative questions often specify variables and expected relationships more explicitly, whereas qualitative questions commonly remain open enough to explore meanings, processes or experiences and may be refined during the study (Barroga & Matanguihan, 2022).

Examples Across Social-Science Fields

Management

Broad topic: Leadership and employee performance.

Weak question: Does leadership affect workers?

Stronger question: What is the relationship between employees’ perceptions of transformational leadership and their self-reported work engagement in selected service-sector organisations?

Why it is stronger: It identifies the relevant constructs, population and type of inquiry without making an unsupported causal claim.

Economics

Broad topic: Education expenditure and development.

Weak question: Does education spending improve a country?

Stronger question: How is government expenditure on secondary education associated with secondary-school completion rates across selected developing economies?

Why it is stronger: It identifies measurable concepts and can potentially be examined using existing secondary data.

Public Policy

Broad topic: Policy implementation.

Weak question: Why are government policies ineffective?

Stronger question: What institutional barriers do local government officials identify in implementing decentralised public-service reforms?

Why it is stronger: It does not assume that the policy is ineffective and identifies the perspective and phenomenon to be investigated.

Education

Broad topic: Digital learning.

Weak question:
Is online learning good for students?

Stronger question: How do university students describe the ways in which blended learning influences their participation in classroom and independent learning activities?

Why it is stronger: It replaces an overly simple evaluative question with a more specific inquiry into experience and participation.

Mixed-Methods Example

A study may need more than one form of evidence.

Question: What is the relationship between participation in digital learning activities and student engagement, and how do students explain their experiences of digital learning?

The quantitative component examines a relationship, while the qualitative component investigates students’ experiences. The researcher would also need to determine how the two forms of evidence will be integrated.

Research Question vs. Research Objective

Research questions and research objectives are closely related but are not the same thing.

A research question states what the study seeks to find out.

A research objective states what the study intends to accomplish.

For example:

The central concepts and intended inquiry should remain aligned.

A common mistake is to write one research question and then create objectives addressing different variables, populations or outcomes.

Good practice: Read each objective beside the corresponding research question and ask whether completing the objective genuinely contributes to answering the question.

Do You Need More Than One Research Question?

Sometimes.

A study may contain:

• one central research question;
• several related subquestions; or
• multiple questions addressing complementary parts of the same research problem.

The important issue is coherence.

Consider:

Central question: How do institutional conditions shape the implementation of digital public services in local governments?

Possible subquestions could examine:

1. What organisational barriers do implementing officials report?
2. What technological constraints do local governments experience?
3. How do officials respond to these constraints?

These questions belong to one coherent inquiry.

By contrast, adding unrelated questions about employee satisfaction, citizen trust, government expenditure and technology adoption could turn one study into several disconnected projects.

Common Mistakes to Avoid

Making the question too broad

Problem: What affects economic development?

Better approach: Identify the particular relationship, outcome, context or mechanism you want to investigate.

Making the question unnecessarily narrow

A question can also become so specific that it loses substantive importance or cannot generate meaningful analysis.

Specificity should serve the research problem—not merely make the title longer.

Assuming the answer

Problem: Why does poor leadership reduce employee motivation?

This assumes that leadership is poor and that it reduces motivation before the study has established either proposition.

More neutral: How do employees perceive the relationship between leadership practices and their work motivation?

Using causal language automatically

Problem: What is the impact of social media use on academic performance?

If the study only measures social-media use and academic performance at one point in time, the design may support an association but not a causal impact.

Possible alternative: What is the relationship between social-media use and academic performance among undergraduate students?

Choosing the method before the question

A researcher may decide, “I want to conduct a survey,” and then search for a question that fits the survey.

The stronger sequence is usually:

Research problem → Research question → Appropriate methodology

Your preferred method should not determine the problem you pretend to investigate.

Choosing a question only because data are available

Existing datasets are valuable, but data availability alone is not a research problem.

Start by identifying a meaningful question, then determine whether existing evidence can answer it appropriately.

Creating too many questions

More research questions do not automatically make a study stronger.

Each additional question introduces additional requirements for evidence, analysis and interpretation.

Ask whether every question is necessary for the central purpose of the study.

Research Question Checklist

Before finalising your question, ask:

Clarity

• Is the question understandable without lengthy explanation?
• Are the main concepts or phenomenon identifiable?
• Have vague terms been removed or defined?

Focus

• Is the scope manageable?
• Does the question address one coherent inquiry?
• Are population, setting or other boundaries included only where they matter?

Researchability

• Can appropriate evidence realistically answer the question?
• Are the required data, participants, documents or records accessible?

Feasibility

• Can the study be completed with the available time, resources and expertise?
• Are ethical and access requirements manageable?

Methodological alignment

• Can the intended design actually answer the question?
• Does the wording avoid claims that the design cannot support?

Relevance

• Does answering the question contribute useful academic, theoretical, policy or practical knowledge?
• Is the question informed by existing literature?

If several answers are no, revise the question before proceeding.