Research questions - developing research questions, what is a good/bad research question...

TL;DR

A good research question must be answerable with available time, resources, and access, and it should be feasible for others to answer as well.

Briefing Cornell Notes

Briefing

A good research question has to be answerable with the time, resources, and access available—and it must avoid assumptions that make the question impossible to verify. Some questions are straightforward because they can be settled with basic data, like whether there are more girls than boys in language classes or whether more females than males attend universities. Those questions are realistic: they don’t require explaining hidden mechanisms, only collecting and comparing counts.

By contrast, questions that ask for a single cause behind a complex pattern often fail because they bundle too many unknown factors into one explanation. “Why do girls perform better in languages than boys?” is flagged as a poor research question because it presumes girls are better language learners in the first place—a claim that may be controversial. Even if the pattern exists, multiple influences could be at work: differences in how learners process language in the brain, cultural and social pressures, classroom motivation, or communication habits that create more opportunities to practice. A study that tries to answer “why girls learn languages better” would struggle to account for all those competing explanations, leaving the research question with limited value.

The criteria for a strong research question come down to three practical tests: size (it should be narrow enough to handle), doability (it should fit the researcher’s resources), and realism (it should be possible to answer). Bad questions are often too broad, not practical, or built on a wrong assumption. A common example is asking “Why do Polish learners enjoy being taught with a task-based methodology?” without first checking whether they actually enjoy it. That framing assumes enjoyment upfront, so the study would need preliminary evidence about enjoyment before it can credibly ask why.

Narrowing a question can feel painful, especially when ambition pushes toward big, important issues. But narrowing improves feasibility and makes the research question more answerable. A personal example illustrates this: a master’s study aimed to test whether having a native English-speaking teacher influences students’ beliefs. Although the researcher tried to control by comparing students taught by native English speakers versus those who were not, other external factors could still shape beliefs. The idea was promising, but the research question needed revision to become truly workable.

Two additional lessons follow. First, research questions can change during a project; collecting data may reveal that the original question no longer matches what the evidence can support. Second, researchers can use multiple research questions and sub-questions to expand coverage while reducing the risk of wrong assumptions. For instance, instead of jumping straight to “why Polish students enjoy task-based teaching,” a better structure asks whether they enjoy it, then—only if enjoyment is present—asks why they enjoy it. That sequencing keeps the study grounded in what can actually be measured and explained.

Cornell Notes

Good research questions must be answerable with available time, resources, and access, and they should avoid built-in assumptions that can’t be verified. Questions that are too broad or that try to explain complex outcomes with a single “why” often fail because multiple factors could be responsible. Narrowing the focus improves feasibility and makes the question more likely to produce meaningful findings. Research questions can also be revised after data collection begins, and using multiple questions or sub-questions helps prevent wrong-assumption problems. A practical example is separating “Do learners enjoy task-based teaching?” from “Why do they enjoy it?” so the study doesn’t assume enjoyment before checking it.

What makes a research question “realistic” versus “not realistic”?

A realistic question can be answered with the researcher’s time, resources, and access, and it can be answered by others as well. Straightforward questions that rely on basic statistics are often realistic (e.g., comparing counts of girls vs. boys in language classes, or comparing male vs. female enrollment in universities). Not realistic questions are often impossible to answer in practice—commonly because they’re too broad, too vague, or require explaining too many unmeasured factors at once.

Why is “Why do girls perform better in languages than boys?” considered a bad research question?

It assumes a contested premise—that girls are better language learners. Even if the pattern is true, the question collapses many possible causes into one explanation, such as brain-based language processing differences, cultural and social influences, motivation, and communication habits that increase practice opportunities. A study would struggle to account for all those factors, so the question becomes hard to answer credibly.

How can a question be “bad” because of a wrong assumption?

A wrong-assumption question presumes the outcome before checking it. For example, asking “Why do Polish learners enjoy being taught with a task-based methodology?” assumes they enjoy it. A better approach is to first ask whether they enjoy it, then ask why—only after enjoyment is established.

Why does narrowing a research question help, even when it feels limiting?

Narrowing reduces scope so the question becomes do-able and answerable with available resources. Broad, ambitious questions often can’t be handled well in a single study. Narrow questions are more feasible to investigate and can still address an important issue, just with a tighter focus.

What lesson comes from the example about native English-speaking teachers and students’ beliefs?

The study idea was to examine whether having a native English-speaking teacher influences students’ beliefs. Even with attempts to control by comparing students taught by native speakers versus non-native speakers, other external factors could still affect beliefs. That mismatch between the question and what can realistically be controlled is why the research question needed to be changed to fit the evidence the study could support.

How can researchers structure questions to avoid wrong-assumption problems?

Using multiple research questions and sub-questions helps. For the task-based teaching example, the main question should test the presence of the phenomenon (“Do Polish students enjoy being taught with task-based methodology?”). Then a sub-question can ask for explanations (“Why do they enjoy it?”). This sequencing prevents the study from assuming the answer before collecting data.

Review Questions

What criteria can be used to judge whether a research question is doable, realistic, and appropriately narrow?
Give an example of how a research question can fail due to either being too broad or being based on a wrong assumption, and propose a better version.
How can changing research questions during data collection improve alignment between the evidence gathered and the conclusions drawn?

Key Points

1
A good research question must be answerable with available time, resources, and access, and it should be feasible for others to answer as well.
2
Questions that rely on simple measurable facts (like counts) are often realistic, while questions requiring explanation of many hidden factors are harder to answer credibly.
3
Avoid framing that assumes the outcome is already true; verify the premise before asking for reasons.
4
Bad research questions are frequently too broad, not practical, or built on incorrect assumptions.
5
Narrowing a question improves feasibility and can strengthen the study rather than weaken it.
6
Research questions can be revised as data collection reveals what the evidence can actually support.
7
Using multiple research questions and sub-questions can reduce wrong-assumption risks and expand what the study can investigate.

Highlights

“Why” questions can be weak when they bundle contested assumptions and multiple unmeasured causes into one explanation.

A wrong-assumption question asks for reasons without first establishing whether the phenomenon exists (e.g., enjoyment).

Narrowing scope is often necessary for feasibility, even when it feels like losing ambition.

Research questions aren’t fixed; they can be reframed when data shows a better fit.

Separating “Do learners enjoy…?” from “Why do they enjoy…?” creates a more defensible study design.

Topics

Research Question Quality
Narrowing Scope
Avoiding Wrong Assumptions
Feasibility and Resources
Sub-Questions