03 · What You Need to Know
The visible problem and the research problem are not necessarily the same thing
Start with what is actually happening
Everyday problems often arrive bundled with explanations.
“Students do not attend consultations because they are not motivated.”
“Staff refuse to use the new system because they resist technology.”
“Patients miss appointments because they forget.”
“Researchers do not share their data because they do not understand open science.”
Each sentence contains an observation and an explanation, but the two can easily become confused. Perhaps consultation schedules conflict with students' classes. Perhaps the new system adds unnecessary steps. Perhaps patients face transportation or childcare barriers. Perhaps researchers understand data sharing but have legitimate concerns about consent, privacy, ownership, or disciplinary norms.
Your first task is therefore to strip the problem back to what you can reasonably observe.
Observation
What appears to be happening without assuming why it happens.
Explanation
A proposed reason or mechanism that might account for the observation.
Research opportunity
An important uncertainty about the observation or explanation that systematic evidence could help resolve.
This separation prevents the study from beginning with its conclusion already embedded in the problem statement.
Ask what you do not know about the problem
Once the problem is described neutrally, interrogate it.
You may not know:
- how common the problem actually is;
- whether it is increasing or decreasing;
- which groups experience it most;
- when or where it occurs;
- what factors are associated with it;
- what mechanism produces it;
- what consequences follow;
- why apparently similar settings experience different outcomes;
- whether an intervention would improve it.
Each uncertainty suggests a different kind of research question. The practical problem gives you the phenomenon. The uncertainty tells you what needs investigation.
A recurring problem is often more informative than an isolated inconvenience
If a printer fails once, you probably need technical support rather than a research protocol. If the same type of breakdown repeatedly occurs across many organizations despite apparently adequate systems, there may be something worth understanding.
Recurrence can suggest that the problem reflects a process rather than an isolated accident. Patterns across people, settings, or time can be particularly informative because they create opportunities to ask what differentiates cases in which the problem occurs from cases in which it does not.
Still, recurrence is not mandatory. Rare events can be consequential enough to warrant investigation. The relevant question is whether understanding the phenomenon would produce knowledge that matters.
Look for variation inside the problem
Variation is one of the richest clues hiding inside everyday problems.
Suppose some instructors adopt a new digital system while others avoid it. Some patients miss appointments repeatedly while others with apparently similar circumstances do not. One branch of an organization implements a new procedure successfully while another struggles.
Instead of asking only “Why is this a problem?”, ask:
What differs between the situations where the problem occurs and those where it does not?
The difference might involve resources, experience, incentives, workload, communication, implementation, accessibility, policy, infrastructure, social relationships, or another mechanism. These possibilities should be investigated rather than assumed, but variation gives you somewhere productive to look.
Workarounds are especially interesting clues
When people repeatedly bypass an official process, researchers sometimes interpret the behavior as resistance or noncompliance. A workaround can instead signal that the formal process does not fit the reality of the work.
Employees may maintain unofficial spreadsheets because the official information system does not support a necessary task. Teachers may use messaging applications because an institutional platform makes communication cumbersome. Clinicians may develop informal routines because formal procedures conflict with workflow.
The workaround itself may therefore reveal a mismatch between system design and actual practice.
If the workaround becomes widespread or consequential, the research question may concern why people adopt it, what need it satisfies, what risks it creates, or what the official process fails to accommodate.
Complaints can contain data, but they are not automatically evidence
Repeated complaints are useful signals. They tell you what people notice, experience, or consider burdensome. They do not necessarily establish the cause, magnitude, or representativeness of the problem.
Ten highly vocal people can create the impression that everyone dislikes a process. Conversely, groups most affected by a problem may be the least able to complain.
Treat complaints as clues requiring investigation. If the problem affects identifiable stakeholders, questions raised by patients, communities, practitioners, or policymakers can help reveal which uncertainties and outcomes matter to them.
Check whether the problem is already understood
A problem can feel new because it is new to you. Before constructing a study, search the literature using both the language people use to describe the practical problem and the terminology researchers use for the underlying phenomenon.
A workplace “workaround” might be studied under technology appropriation, implementation, workflow adaptation, shadow systems, or related concepts. Students “giving up” might connect to research on persistence, self-efficacy, cognitive load, motivation, help-seeking, or belonging depending on what is actually happening.
Relevant evidence may also come from another discipline. Everyday problems rarely respect departmental boundaries with the same enthusiasm universities do.
Do not assume that local novelty means scientific novelty
You may be the first person to study a problem in your institution, organization, city, or profession. That can matter when local evidence is needed for a local decision, but it does not automatically create a broader research contribution.
Ask whether the local setting introduces a condition that could change what existing research would predict. If not, the project may still be valuable as evaluation, quality improvement, needs assessment, or organizational inquiry without needing to be presented as a major gap in scientific knowledge.
The label matters less than intellectual honesty about what the project is designed to accomplish.
Some everyday problems need management, not research
Suppose an office discovers that employees cannot submit a form because the hyperlink is broken. The cause is known, the solution is obvious, and fixing the link resolves the problem. Research would add little.
Now suppose forms are technically accessible, reminders are sent, support is available, yet completion remains consistently low and varies dramatically among employee groups. The cause is uncertain, plausible explanations differ, and the problem affects important organizational decisions. Research becomes more defensible.
| Mostly a troubleshooting problem |
Potential research opportunity |
| The cause is already clear |
The cause remains uncertain or contested |
| The solution is obvious and established |
Several plausible solutions or explanations exist |
| The event is isolated and inconsequential |
The problem recurs or has meaningful consequences |
| Evidence already provides a sufficient answer |
Existing evidence does not adequately resolve the relevant uncertainty |
| Fixing the immediate process is sufficient |
Understanding the problem could inform other cases, settings, decisions, or theory |
Do not confuse quality improvement with research automatically
Projects designed primarily to improve a local process and research intended to produce generalizable knowledge can overlap in methods, but their purposes may differ. Regulatory and institutional definitions also vary, particularly in health and organizational settings.
If your project involves human participants, identifiable information, institutional records, interventions, or potentially sensitive data, determine the applicable ethics and governance requirements through the appropriate institutional process rather than deciding for yourself that an activity is “just improvement.”
Whether a problem originates locally does not remove ethical responsibilities.
A real-world problem can lead to several different studies
One problem does not imply one inevitable research question.
Consider low participation in a student-support service. You might investigate prevalence and patterns of nonuse, barriers to access, students' understanding of the service, differences among student groups, referral processes, implementation across departments, or the effect of a redesigned service.
These are distinct questions requiring different designs.
This is why recognizing a real-world problem as a possible research topic is only an early step. You still need to identify which uncertainty within that problem deserves investigation.
The strongest clue is often an assumption nobody has checked
Everyday systems run on assumptions.
Students will read the instructions. Patients will understand the reminder. Staff will use the new software. Managers will act on the dashboard. Users will prefer the more efficient option. Practitioners will implement the guideline as designed.
When reality repeatedly violates one of these assumptions, you may have found something interesting.
Ask:
What did this process assume people would do, know, prefer, or have access to, and is that assumption actually true?
A study that examines a hidden assumption can produce more transferable knowledge than one that merely documents the visible failure.
Watch Out
Do not build a study merely to confirm your favorite explanation for a frustrating problem. Everyday experience is excellent at generating hypotheses and equally capable of making one explanation feel obvious before the evidence has had a chance to disagree.