A research problem is a specific, evidence-supported gap, contradiction, difficulty, or unresolved issue that a study investigates. It explains what is not known or not working, who or what is affected, why the issue matters, and the boundaries within which the researcher will examine it.
Introduction
Every research project needs more than an interesting subject. It needs a clearly defined problem that explains why systematic investigation is necessary.
A student may be interested in social media, climate change, artificial intelligence, employee motivation or public health. These are research topics, not research problems. A research problem emerges only when the researcher identifies a specific uncertainty, contradiction, limitation, unmet need or practical difficulty within the broader topic.
This guide explains what a research problem is, how it differs from related concepts, where research problems come from, how to formulate one, and how the problem should align with the research questions, objectives, design and analysis.
Key Takeaways
- A research problem identifies a specific issue, gap, contradiction or difficulty requiring systematic investigation.
- It must be supported by credible evidence rather than personal opinion alone.
- A topic states the general subject; a problem explains what is unknown, inadequate or unresolved within that subject.
- A strong problem is significant, focused, researchable, feasible, ethical and appropriately original.
- The research problem should logically lead to the research questions, objectives and methodological choices.
- AI tools may assist with searching and organising literature, but they cannot independently prove that a problem is original.
What Is a Research Problem?
A research problem is a clearly bounded issue that creates a justified need for research. It may involve missing knowledge, conflicting findings, an inadequate theory, a poorly understood experience, an ineffective practice, a measurement weakness or a real-world condition that has not been adequately investigated.
The word problem does not always mean that something harmful has happened. In research, it can also refer to:
- An unanswered question.
- An inconsistency between studies.
- A phenomenon that is insufficiently understood.
- A theory that has not been tested in a relevant context.
- A method that produces unreliable or incomplete results.
- A policy or intervention whose effects are uncertain.
- A population whose experiences have been overlooked.
A research problem therefore explains the intellectual or practical reason for conducting a study.
Simple example
Topic: Online learning
Research problem: Universities have expanded asynchronous online courses, but existing evidence does not adequately explain how limited real-time interaction affects first-year students’ sense of belonging and persistence.
The topic identifies the field. The problem identifies what is not adequately understood, the relevant setting and the potential consequence.
Why Is the Research Problem Important?
A research problem gives the study direction, justification and coherence.
It establishes the purpose of the study
The problem explains why the study deserves to be conducted. Without it, a project may collect information without making a meaningful contribution.
It determines the scope
A well-defined problem identifies the population, setting, concepts, variables or period that the study will examine. This prevents the project from becoming unmanageably broad.
It guides the research questions and objectives
Research questions translate the problem into answerable inquiries. Objectives state what the researcher will do to answer those questions.
It influences the methodology
Different problems require different forms of evidence. A prevalence problem may require a survey, an experience problem may require interviews, and a causal problem may require an experimental or strong quasi-experimental design.
It supports evaluation
Supervisors, ethics committees, reviewers and funders use the problem to assess whether the proposed study is relevant, feasible and methodologically coherent.
It helps communicate the contribution
The eventual contribution should respond directly to the original problem. Readers need to understand what was uncertain before the study and what became clearer after it.
Research Topic, Gap, Problem, Statement, Question, Objective and Hypothesis
These concepts are connected but not interchangeable.
| Concept | Meaning | Example |
|---|---|---|
| Research topic | The broad subject area | Social media and adolescents |
| Research gap | Something absent, limited, inconsistent or inadequate in existing knowledge | Few longitudinal studies have examined passive social-media use and sleep among younger adolescents |
| Research problem | The specific evidence-supported issue that justifies investigation | It remains unclear whether sustained passive social-media use predicts changes in sleep quality among adolescents aged 13–15 |
| Problem statement | The written explanation of the problem, context, evidence and significance | A paragraph explaining the uncertainty, affected population, evidence and consequences |
| Research question | The precise question the study will answer | How is passive social-media use associated with changes in sleep quality among adolescents aged 13–15 over one school year? |
| Research objective | A statement of what the study will do | To estimate the relationship between passive social-media use and changes in sleep quality |
| Hypothesis | A testable prediction used when appropriate | Higher passive social-media use will be associated with a greater decline in sleep quality |
Research gap versus research problem
A research gap is an absence, weakness or inconsistency in existing knowledge. The research problem is the larger justification for investigating that gap.
For example, “few studies have examined rural students” is a gap claim. It becomes a defensible research problem only when the researcher explains:
- Why the rural context may produce meaningfully different results.
- Which existing findings may not transfer to that setting.
- Who is affected.
- What theoretical, practical or policy consequence follows from the uncertainty.
Not every gap is important enough to support a study.
Research problem versus problem statement
The research problem is the underlying issue. The problem statement is the written passage that communicates it.
A problem may require extensive reading and conceptual development. Its final statement may still be expressed in one focused paragraph.
Research problem versus research question
The problem is normally written as a declarative explanation of what is unknown or inadequate. The research question converts that problem into an interrogative form that can guide data collection and analysis.
Characteristics of a Good Research Problem
A strong research problem should be:
Clear
Readers should understand the central issue without decoding unnecessary jargon.
Specific
The problem should identify relevant boundaries such as the population, setting, concepts, variables, time frame or unit of analysis.
Evidence-supported
The researcher should demonstrate the problem using credible literature, data, reports, observations or stakeholder evidence.
Researchable
The issue must be open to systematic investigation using appropriate evidence. Purely moral, metaphysical or personal claims may not be empirically researchable unless reframed.
Significant
The answer should matter to a scholarly, professional, community or policy audience.
Feasible
The project must be possible within the available time, expertise, budget, access and data constraints.
Ethical
The problem should be investigable without imposing unjustifiable harm, coercion, privacy invasion or inequitable burdens.
Appropriately original
Originality does not always require studying a completely untouched subject. A study may contribute through a new theory, method, context, dataset, comparison, replication, interpretation or synthesis.
Aligned
The problem, question, objectives, design, evidence and analysis should form a logical chain.
Main Types of Research Problems
No single taxonomy applies universally across all disciplines. The following categories describe common sources of researchable problems.
| Type | Central issue | Example |
|---|---|---|
| Knowledge or empirical problem | Evidence is absent, limited or inconclusive | Little is known about how refugee entrepreneurs obtain informal business support in small European cities |
| Contradiction problem | Credible studies produce conflicting findings | Studies disagree about whether hybrid working improves team innovation |
| Theoretical problem | A theory is incomplete, contested or insufficiently tested | Existing technology-acceptance models may not explain refusal to use mandatory workplace AI |
| Methodological problem | Existing methods or measures are inadequate | Common self-report measures may not accurately capture short periods of digital distraction |
| Measurement problem | A construct is inconsistently defined or operationalised | Researchers use incompatible definitions of student engagement, preventing meaningful comparison |
| Population or contextual problem | Evidence may not transfer to a neglected population or setting | Most evidence on telehealth access comes from urban populations, leaving remote island communities underexamined |
| Applied or practical problem | A real-world process, service or outcome is not working adequately | A hospital’s appointment-reminder system has not reduced missed visits among patients requiring interpreters |
| Policy or evaluation problem | The effects or implementation of a policy are uncertain | It is unclear whether a new attendance policy reduces absenteeism or disproportionately penalises students with chronic illness |
| Design or engineering problem | Available solutions do not meet defined requirements | Existing low-cost air-quality sensors lose accuracy under high-humidity conditions |
| Equity or representation problem | Research or practice overlooks unequal effects or missing voices | Automated recruitment tools may produce different screening outcomes across demographic groups, but local audit evidence is unavailable |
| Replication or robustness problem | An influential finding has not been independently reproduced | A widely used classroom intervention has not been replicated across schools with different staffing levels |
| Implementation problem | An evidence-supported intervention is not adopted or sustained effectively | Clinicians accept a guideline in principle but rarely apply it during high-pressure emergency decisions |
Practical and theoretical problems can overlap
A study may address both a practical challenge and a theoretical uncertainty. For example, research on low adoption of a medical application could improve implementation while also testing whether an existing technology-adoption theory explains clinician behaviour.
The researcher should identify the primary problem so that the project remains focused.
Where Do Research Problems Come From?
Research problems may emerge from several sources.
Existing literature
Journal articles, scholarly books and reviews may reveal:
- Unanswered questions.
- Contradictory findings.
- Weak measures.
- Untested assumptions.
- Inconsistent definitions.
- Underrepresented populations.
- Limitations of previous designs.
- Recommendations for replication or extension.
The “limitations” and “future research” sections can generate ideas, but they should not be accepted uncritically. Authors may overlook studies or recommend projects that are no longer necessary.
Systematic and scoping reviews
Evidence syntheses can reveal where knowledge is abundant, weak, inconsistent or absent. They are particularly useful for determining whether an apparent gap is real.
Theory
A researcher may identify assumptions that have not been adequately examined or a theory that does not explain a new phenomenon. Problematization goes beyond locating an empty space in the literature by questioning the assumptions on which existing explanations depend (Alvesson & Sandberg, 2011).
Professional practice
Teachers, clinicians, engineers, managers, social workers and policymakers encounter recurring problems that may not be adequately represented in published research.
Observation and preliminary data
Institutional records, pilot studies, audits, public datasets and field observations may reveal unexpected patterns.
Stakeholder and community priorities
Patients, students, employees and community organisations may identify problems that researchers have overlooked. Their involvement can improve relevance, although consultation should not be treated as a substitute for literature validation.
Social, technological or policy change
New technologies, laws, crises and practices can create new research needs. Generative AI, remote work and climate adaptation are examples of developments that may produce new questions or alter older ones.
Failed or incomplete solutions
A problem may arise when an established intervention works inconsistently, produces unintended consequences or cannot be implemented in a particular context.
How to Identify and Formulate a Research Problem
Step 1: Select a broad area of interest
Begin with a subject that is relevant to your discipline, programme or professional setting.
Examples include:
- Student retention.
- Renewable energy.
- Telemedicine.
- Workplace automation.
- Language learning.
- Cybersecurity behaviour.
At this stage, do not try to write the final research question.
Step 2: Conduct a preliminary evidence search
Search recent reviews, major studies, foundational theories, policy documents and relevant datasets.
Use multiple databases where possible. A single search engine may miss discipline-specific journals, alternative terminology or non-English evidence.
Record:
- Search terms and synonyms.
- Databases used.
- Date ranges.
- Frequently cited studies.
- Recent reviews.
- Important disagreements.
- Populations and settings studied.
- Measures and methods used.
Step 3: Map what is known
Summarise the strongest conclusions that the existing evidence supports.
Ask:
- What findings are reasonably established?
- Which theories dominate the field?
- What methods are commonly used?
- Which populations and settings receive the most attention?
- Which definitions or measures are accepted or disputed?
A research problem cannot be defined clearly until the existing knowledge baseline is understood.
Step 4: Identify a meaningful deficiency or tension
Look for something that is:
- Unknown.
- Inconsistently explained.
- Poorly measured.
- Insufficiently tested.
- Ineffectively implemented.
- Contextually uncertain.
- Theoretically disputed.
- Practically inadequate.
Avoid assuming that “no study in my city” is sufficient. Explain why the local setting creates a meaningful scientific or practical difference.
Step 5: Verify that the problem still exists
A statement in an older article may have been resolved by newer research.
Validate the problem by:
- Searching recent publications.
- Checking systematic or scoping reviews.
- Following studies that cite major papers.
- Searching alternative terminology.
- Consulting a librarian, supervisor or subject specialist.
- Reviewing relevant policies, datasets or professional reports.
- Looking for registered or ongoing studies where relevant.
Novelty is a conclusion based on a search process, not an assumption.
Step 6: Explain why the problem matters
Identify the consequence of leaving the issue unresolved.
The significance may be:
- Theoretical: a theory remains incomplete.
- Empirical: evidence remains uncertain.
- Methodological: findings cannot be compared reliably.
- Practical: a service or intervention remains ineffective.
- Social: a population continues to experience harm or exclusion.
- Policy-related: decision-makers lack reliable evidence.
- Ethical: risks or unequal effects remain unexamined.
Step 7: Define the boundaries
Specify what the study will and will not address.
Possible boundaries include:
- Population.
- Geographic setting.
- Institution or industry.
- Period.
- Variables or concepts.
- Technology or intervention.
- Unit of analysis.
- Type of evidence.
- Theoretical perspective.
Boundaries should make the study manageable without making it trivial.
Step 8: Test feasibility and ethics
Ask:
- Can the relevant participants, documents or data be accessed?
- Is the sample or evidence source realistic?
- Can the key concepts be measured or explored?
- Does the researcher possess the required skills?
- Can the project be completed within the available time?
- Will ethics approval be required?
- Could the study create harm, stigma, privacy risks or unfair burdens?
- Is a safer or less intrusive method available?
The FINER criteria—feasible, interesting, novel, ethical and relevant—are widely used for evaluating research questions and can also help assess a proposed problem (Farrugia et al., 2010).
Step 9: Write the problem in one sentence
Use a working sentence such as:
Although [what is already known or being done], it remains unclear or inadequate [specific gap, contradiction or difficulty] among [population or setting], which matters because [consequence].
Example:
Although universities increasingly use AI-supported writing feedback, it remains unclear how students with limited academic-English proficiency interpret and act on that feedback, potentially limiting equitable learning benefits.
Step 10: Align the problem with the study
Check whether the problem leads logically to:
- The research question.
- The purpose or aim.
- The objectives.
- The theoretical or conceptual framework.
- The design.
- The data required.
- The analysis.
- The expected contribution.
If the method cannot answer the question created by the problem, revise either the problem, question or design.
How to Write a Research Problem Statement
A problem statement is a concise, evidence-supported explanation of the issue that the study will address.
There is no universal word count. A short proposal may use one focused paragraph, while a dissertation or funding application may develop the problem over several paragraphs. Follow the requirements of the institution, discipline, journal or funder.
Core components
A strong problem statement normally contains:
- Context: What is already known or occurring?
- Evidence: What research or data establishes the situation?
- Deficiency: What is missing, inconsistent, ineffective or unresolved?
- Scope: Who or what is affected, and where?
- Consequence: Why does the deficiency matter?
- Research direction: What will the study investigate?
Reusable template
Context: [Describe the relevant situation and established knowledge.]
Deficiency: However, [identify the precise gap, contradiction, limitation or practical difficulty].
Scope: This problem concerns [population, setting, variables, concepts or period].
Significance: Without better evidence, [explain the theoretical, practical, social or policy consequence].
Direction: Therefore, this study will [state the broad research purpose without promising a predetermined result].
Example problem statement
Universities increasingly use automated feedback systems to support academic writing, and previous research suggests that rapid formative feedback can improve revision activity. However, limited evidence explains how multilingual first-year students interpret automated comments, decide which suggestions to accept and distinguish language correction from substantive academic guidance. This uncertainty may cause institutions to overestimate the educational value and equity of automated feedback. The proposed study will therefore examine how multilingual first-year students understand and use AI-generated writing feedback during a semester-long academic-writing course.
What not to include
Avoid:
- Unsupported claims.
- Emotional or exaggerated language.
- A predetermined conclusion.
- An entire literature review.
- Detailed method descriptions.
- A list of unrelated gaps.
- Claims that the study will completely “solve” a complex social problem.
- Citations generated by AI that have not been verified.
Research Problem Examples
The examples below illustrate possible formulations. They are not claims that the stated gaps currently exist; a researcher would need to verify each one through a current literature search.
| Discipline | Example research problem |
|---|---|
| Education | Schools have adopted adaptive-learning platforms, but it remains unclear whether the feedback produced by these systems supports self-regulated learning among students with low prior attainment. |
| Psychology | Existing studies link notification frequency with distraction, but the mechanisms through which unpredictable notifications affect sustained attention remain insufficiently understood. |
| Public health | Vaccination services are available in the target district, yet uptake remains uneven and existing administrative data do not explain how transport, trust and appointment flexibility interact. |
| Nursing | Burnout among emergency nurses is well documented, but less is known about how informal peer-support practices influence recovery after ethically distressing incidents. |
| Business | Small firms increasingly adopt generative-AI tools, but evidence is limited on how informal governance practices affect errors, confidentiality and employee accountability. |
| Computer science | Current phishing-detection models perform well on benchmark datasets but lose accuracy when messages combine multiple languages and culturally specific cues. |
| Engineering | Low-cost environmental sensors enable wider monitoring, but their accuracy under rapid temperature and humidity changes remains inadequate for local regulatory decisions. |
| Environmental science | Urban tree-planting programmes report total canopy growth, but limited evidence shows whether cooling benefits are distributed equitably across neighbourhoods with different housing densities. |
| Sociology | Remote work has changed workplace interaction, yet the effects of reduced informal contact on the professional inclusion of newly hired employees remain poorly understood. |
| Public policy | A new digital-benefits application has reduced processing time overall, but its effects on applicants with limited digital literacy have not been adequately evaluated. |
| History | Historians agree that wartime broadcasting influenced public communication, but disagree about the extent to which regional stations shaped local political identity. |
| Linguistics | Automated speaking assessments are increasingly used, yet their treatment of legitimate regional accent variation remains insufficiently transparent. |
Weak and Strong Research Problems
Example 1: Too broad
Weak: Technology affects education.
Why it is weak: It does not identify a technology, population, outcome, setting, gap or significance.
Stronger: Although secondary schools increasingly use automated formative quizzes, it remains unclear whether repeated algorithmic difficulty adjustments improve conceptual retention among students who begin a course with low subject knowledge.
Example 2: A policy opinion
Weak: Universities should ban smartphones from classrooms.
Why it is weak: It assumes a solution before investigating the issue.
Stronger: Evidence remains inconsistent regarding how different classroom smartphone policies affect attention, participation and accessibility for university students.
Example 3: An unsupported local gap
Weak: No one has studied employee motivation in my company.
Why it is weak: A missing company-specific study does not automatically establish scholarly or practical importance.
Stronger: Employee exit interviews indicate persistent dissatisfaction despite recent financial incentives, but the organisation lacks evidence about how supervisory autonomy, workload predictability and promotion transparency influence intention to leave.
Example 4: A question instead of a problem
Weak: Does social media cause anxiety?
Why it is weak: It is broad, uses causal language without specifying a design and does not establish the evidence gap.
Stronger: Longitudinal evidence remains inconclusive about whether changes in passive social-media use precede changes in anxiety symptoms among younger adolescents.
Quantitative, Qualitative and Mixed-Methods Problem Framing
The research approach should respond to the nature of the problem rather than being selected only because the researcher prefers a particular method.
| Approach | Suitable problem focus | Typical wording | Example |
|---|---|---|---|
| Quantitative | Prevalence, differences, associations, prediction, effects or measurable patterns | Extent, relationship, difference, effect, prediction | It is unclear whether attendance at supplemental instruction predicts course completion after controlling for prior attainment |
| Qualitative | Experiences, meanings, perceptions, processes or poorly understood phenomena | How, why, experience, meaning, process | Little is known about how first-generation students interpret requests for academic help |
| Mixed methods | A problem requiring both measurable patterns and contextual explanation | Extent plus how or why | Retention rates differ across programmes, but administrative data do not explain how students experience the support systems associated with those differences |
Quantitative problem formulation
A quantitative problem should identify measurable variables or outcomes without overstating causality.
A cross-sectional association does not by itself demonstrate that one variable causes another. The wording of the problem should match the strength of the proposed design.
Qualitative problem formulation
A qualitative problem may focus on an experience or process that has not been adequately understood. It does not always need predetermined variables or a formal hypothesis.
The concepts and scope should still be sufficiently clear to guide participant selection and data collection.
Mixed-methods problem formulation
A mixed-methods problem should explain why one form of data is insufficient. The researcher might need quantitative data to establish a pattern and qualitative data to explain how or why the pattern occurs.
Using two methods without an integration rationale does not create a strong mixed-methods study.
How the Research Problem Influences the Research Design
| Problem purpose | Typical question | Potential designs |
|---|---|---|
| Describe | What is happening, how often or to whom? | Survey, descriptive study, content analysis, observational study |
| Explore | How is a poorly understood phenomenon experienced? | Interviews, focus groups, ethnography, case study |
| Explain association | How are variables related? | Correlational, cross-sectional or longitudinal design |
| Estimate causal effect | What difference does an intervention or exposure make? | Randomised experiment, natural experiment, quasi-experiment |
| Predict | Which factors forecast an outcome? | Cohort design, predictive modelling, validation study |
| Evaluate | How well does a programme or policy work, for whom and under what conditions? | Process, outcome, impact or realist evaluation |
| Develop theory | What process or mechanism explains the phenomenon? | Grounded theory, theory-building case study |
| Design a solution | How can a defined need be met? | Design science, engineering design, prototyping and testing |
| Synthesise evidence | What does the total evidence show? | Systematic review, meta-analysis, qualitative synthesis or scoping review |
These associations are guides rather than automatic rules. The final design also depends on ethics, access, data quality, theory and disciplinary conventions.
How to Evaluate a Proposed Research Problem
Use the following checklist before finalising the problem.
Evidence
- Have I shown that the problem exists?
- Have I searched recent as well as foundational literature?
- Have I checked alternative terminology?
- Have I distinguished an actual gap from a weak search?
- Are the main claims supported by credible sources?
Significance
- Who needs the answer?
- What theoretical, empirical, practical or policy consequence follows?
- Would addressing the problem change understanding, decisions or practice?
- Is the contribution more meaningful than simply repeating a study in a new location?
Scope
- Is the population defined?
- Is the setting clear?
- Are the main concepts or variables identifiable?
- Is the time frame appropriate?
- Have exclusions and boundaries been considered?
Researchability
- Can the problem be investigated systematically?
- Can relevant evidence be collected or accessed?
- Are the concepts observable, measurable or meaningfully interpretable?
- Can the proposed questions be answered rather than merely discussed?
Feasibility
- Is the project manageable within the available time and budget?
- Is the sample or dataset accessible?
- Does the researcher have the necessary methodological expertise?
- Are equipment, software, language or travel requirements realistic?
Ethics
- Are participant risks proportionate to the value of the research?
- Could the wording stigmatise a population?
- Are privacy, consent and data-security issues manageable?
- Have affected communities been represented appropriately?
- Could the study worsen an existing inequality?
Alignment
- Does each research question respond directly to the problem?
- Does each objective support a question?
- Can the design generate the required evidence?
- Does the planned analysis match the question?
- Would the expected findings address the original problem?
Common Mistakes When Formulating a Research Problem
Choosing a topic instead of a problem
“Artificial intelligence in education” is a topic. The researcher must identify a specific uncertainty, difficulty or contradiction within it.
Claiming that no research exists without sufficient searching
Absolute novelty claims are difficult to defend. Prefer cautious language such as “limited evidence was identified” and document the search process.
Treating novelty as the only criterion
An unstudied issue may be unimportant, unethical or impossible to investigate. Significance and feasibility matter alongside novelty.
Starting with a preferred method
A researcher should not create a problem merely to justify using a survey, experiment or machine-learning model. Research should be problem-driven rather than method-driven.
Prescribing the answer
“Schools must replace teachers with AI tutors” is a position, not a neutral research problem.
Using causal language with a weak design
Words such as effect, impact and causes imply stronger evidence than association or relationship. The language should match the planned design.
Combining multiple unrelated problems
A dissertation cannot normally address every weakness in a field. Select one central problem and treat additional issues as context or future research.
Assuming a new context automatically creates a contribution
A geographical replication is valuable only when there is a reason to expect context to influence the result or when local evidence is necessary for decisions.
Confusing social importance with researchability
A severe social issue may still be too broad for one study. The researcher must identify an investigable component.
Writing the problem before reviewing the literature
Initial experience may suggest a problem, but literature and stakeholder evidence are needed to refine and validate it.
Digital Research Tools and Responsible AI Use
Modern tools can make problem identification more systematic, but no tool can replace scholarly judgement.
Literature databases
Depending on the discipline, researchers may use:
- Google Scholar.
- Scopus.
- Web of Science.
- PubMed.
- ERIC.
- PsycINFO.
- IEEE Xplore.
- Subject-specific indexes.
- Institutional repositories.
- Government and international-organisation databases.
Searching more than one source reduces the chance that an apparent gap is caused by incomplete coverage.
Reference managers
Zotero, EndNote and Mendeley can help researchers organise sources, notes, tags and citations. A useful tagging system might include:
- Established finding.
- Contradiction.
- Limitation.
- Population gap.
- Method problem.
- Theory issue.
- Future research.
- Potential significance.
Citation-network and evidence-mapping tools
Citation mapping can help identify influential papers, related clusters and later studies that cite a foundational work. These tools support discovery but should not replace structured database searches.
How AI can assist
Generative AI may help with:
- Suggesting search-term synonyms.
- Rephrasing a broad interest into several candidate problem formulations.
- Grouping researcher-supplied abstracts by theme.
- Testing whether a draft problem statement is understandable.
- Identifying missing components in a researcher-created template.
- Generating alternative non-causal wording.
- Creating questions for discussion with a supervisor.
What AI should not be trusted to do independently
AI should not be relied upon to:
- Prove that a research gap exists.
- Guarantee originality.
- Produce references without verification.
- Summarise studies that the researcher has not checked.
- Decide whether a problem is ethically acceptable.
- Select a methodology without disciplinary judgement.
- Handle confidential proposals or unpublished data in an unapproved system.
- Be listed as an author.
Generative systems may return plausible but inaccurate citations or overlook relevant evidence. UNESCO guidance emphasises human-centred, transparent and accountable use of generative AI in education and research. Publication policies also increasingly require authors to remain responsible for accuracy and disclose relevant AI assistance (UNESCO, 2023; ICMJE, 2026).
Researchers should check their university, funder and target journal policies because disclosure requirements differ.
Ethical and Practical Considerations
Avoid deficit-based framing
A problem statement should not portray a community as inherently deficient when structural, institutional or historical conditions may explain the observed outcome.
Instead of writing:
Rural patients fail to use digital services.
Consider:
Existing digital services may not adequately account for connectivity, language and accessibility conditions in rural communities.
Separate evidence from interpretation
Administrative records may show that an outcome exists, but they may not explain why it exists. The problem statement should distinguish observed evidence from proposed explanations.
Do not exaggerate urgency
Terms such as crisis, devastating or unprecedented require evidence. Precise language is more credible than dramatic language.
Consider who defines the problem
A university, company or policymaker may define a situation differently from the people affected by it. Stakeholder involvement can reveal assumptions and unintended consequences.
Protect confidentiality
Do not upload unpublished proposals, identifiable participant information, confidential institutional records or peer-review manuscripts to external AI systems without appropriate permission and safeguards.
Do not promise complete resolution
A single study normally provides a bounded contribution. It may clarify, estimate, explain, test or evaluate part of a larger problem rather than solve it permanently.
Can a Research Problem Change During a Study?
Yes. A research problem may be refined as the researcher reads more literature, gains access to data, conducts pilot work or receives ethical and supervisory feedback.
In exploratory and qualitative research, early evidence may reveal that the original framing was incomplete. In quantitative and confirmatory research, major changes after data examination can introduce bias and should be documented transparently.
A legitimate refinement should be distinguished from changing the problem merely to make unexpected findings appear predicted. Where preregistration or a formal protocol applies, amendments should be recorded and justified.
Research Problem Worksheet
Complete the following statements:
Broad topic:
This project concerns ____________________________________.
Established knowledge:
Existing research generally shows _________________________.
Evidence deficiency or practical difficulty:
However, it remains unclear/inconsistent/inadequate whether or how ____________________________________.
Population or setting:
This problem affects or concerns __________________________.
Significance:
The problem matters because ______________________________.
Scope:
The study will focus on ___________________________________ and will not attempt to address ___________________________________.
Research direction:
The study will investigate ________________________________.
Provisional research question:
_________________________________________________________?
Feasibility evidence:
I can access _____________________________________________.
Ethical consideration:
The main ethical issue is _________________________________, which will be addressed by ___________________________________.
Conclusion
A research problem is the evidence-supported reason a study needs to be conducted. It identifies what is unknown, inconsistent, inadequate or unresolved and explains why that deficiency matters within a defined context.
The strongest research problems are not merely broad topics or empty spaces in the literature. They are significant, researchable, feasible, ethical and logically connected to the research questions, objectives, methods and intended contribution.
Frequently Asked Questions
What is a research problem in simple words?
A research problem is a specific issue or unanswered matter that requires systematic investigation. It explains what is missing, uncertain or not working and why obtaining better evidence would be valuable.
What is an example of a research problem?
An example is: “Universities increasingly use automated writing feedback, but limited evidence explains how multilingual first-year students interpret and apply that feedback.” The statement identifies a setting, a population and an unresolved issue.
What are the main types of research problems?
Common types include empirical, theoretical, methodological, measurement, practical, policy, contextual, contradiction, replication, implementation and equity-related problems. These categories can overlap.
What is the difference between a research gap and a research problem?
A research gap is a specific weakness or absence in existing knowledge. A research problem explains why that gap or difficulty deserves investigation, who or what it affects and what consequences follow from leaving it unresolved.
What is the difference between a research problem and a research question?
The research problem describes and justifies the issue. The research question converts that issue into a precise question that the study will answer.
How long should a research problem statement be?
There is no universal length. It may be one focused paragraph in a short proposal or several paragraphs in a dissertation. It should include enough evidence to establish the context, deficiency, scope and significance without becoming a complete literature review.
Does every research problem require a hypothesis?
No. Hypotheses are mainly used when a study tests predicted relationships or differences. Exploratory, qualitative, descriptive and some mixed-methods studies may use research questions without formal hypotheses.
Can one study have more than one research problem?
A project may contain related subproblems, but it should normally have one central problem. Multiple unrelated problems weaken coherence and may make the project unmanageable.
Can a research problem be based on personal experience?
Personal or professional experience can generate an initial idea, but the problem should be validated using literature, data, stakeholder evidence or other credible sources.
Can AI identify a research problem?
AI can suggest terminology, organise researcher-supplied information and help improve wording. It cannot independently verify novelty, guarantee accurate references or decide whether the problem is significant, feasible or ethical. Human verification remains essential.
Suggested authoritative sources
The following are appropriate for an APA 7 reference list. Edition and journal requirements should still be checked against the version actually consulted.
- Alvesson, M., & Sandberg, J. (2011). Generating research questions through problematization. Academy of Management Review, 36(2), 247–271. doi:10.5465/amr.2009.0188.
- Appelbaum, M., Cooper, H., Kline, R. B., Mayo-Wilson, E., Nezu, A. M., & Rao, S. M. (2018). Journal article reporting standards for quantitative research in psychology: The APA Publications and Communications Board task force report. American Psychologist, 73(1), 3–25.
- Creswell, J. W., & Creswell, J. D. (2023). Research design: Qualitative, quantitative, and mixed methods approaches (6th ed.). SAGE.
- Ellis, T. J., & Levy, Y. (2008). Framework of problem-based research: A guide for novice researchers on the development of a research-worthy problem. Informing Science: The International Journal of an Emerging Transdiscipline, 11, 17–33. doi:10.28945/3288.
- Farrugia, P., Petrisor, B. A., Farrokhyar, F., & Bhandari, M. (2010). Research questions, hypotheses and objectives. Canadian Journal of Surgery, 53(4), 278–281.
- Levitt, H. M., Bamberg, M., Creswell, J. W., Frost, D. M., Josselson, R., & Suárez-Orozco, C. (2018). Journal article reporting standards for qualitative primary, qualitative meta-analytic, and mixed methods research in psychology. American Psychologist, 73(1), 26–46. doi:10.1037/amp0000151.
- American Psychological Association. (n.d.). Reporting your research: How to use APA Style Journal Article Reporting Standards. s.
- EQUATOR Network. (n.d.). Selecting the appropriate reporting guideline for your article.
- International Committee of Medical Journal Editors. (2026). Use of artificial intelligence by authors.
- UNESCO. (2023). Guidance for generative AI in education and research.
