Research Begins Before Methodology

Research students often arrive at methodology remarkably early.

They ask:

Should I use a questionnaire?

How many interviews do I need?

Should the research be qualitative or quantitative?

Which statistical test should I use?

These are important questions.

But frequently, they are being asked before another set of questions has been answered:

What exactly am I trying to understand?

Why does it need to be understood?

What kind of evidence could actually help me understand it?

A method is meaningful only in relation to a problem.

Research therefore begins considerably earlier than methodology.

The temptation to begin with the method

Methods feel concrete.

A questionnaire can be designed.

An interview can be scheduled.

Data can be entered into software.

A statistical test can be named in a proposal.

These activities create the reassuring impression that research has begun.

But activity is not necessarily progress.

Consider a student who says:

“I am doing a quantitative study using a questionnaire.”

We still know surprisingly little about the research.

What is the problem?

What does the researcher want to explain?

What variables matter?

Why should those variables be related?

Why is a questionnaire capable of capturing them?

What population matters?

What would constitute a convincing answer?

Until those questions are addressed, “quantitative study using a questionnaire” describes a technique—not a research design.

Research begins with uncertainty

At its core, research begins because something is not adequately understood.

Perhaps we do not know whether two phenomena are related.

Perhaps existing evidence contradicts itself.

Perhaps we know that something happens but do not understand why.

Perhaps a theory explains one context but performs poorly in another.

Perhaps people experience something in ways that existing categories fail to capture.

Perhaps a policy is being implemented without sufficient evidence about its consequences.

That uncertainty is intellectually more important than the instrument eventually used to investigate it.

The researcher’s first task is therefore not:

Choose a method.

It is:

Understand the uncertainty.

A topic is only the beginning

Suppose a student says:

“My research is about social media.”

That identifies a broad territory.

It does not yet identify research.

We could study:

social media and political participation,

social media and academic distraction,

misinformation,

identity formation,

consumer behaviour,

mental health,

language change,

online communities,

or hundreds of other questions.

Narrowing the topic helps, but even:

“Social media and university students”

still tells us very little about the intellectual purpose of the study.

A researcher has to move beyond what the subject is toward what is not understood about it.

That movement—from topic to problem—is one of the most important transitions in the entire research process.

The order matters.

When we reverse it, research becomes method-driven.

When we preserve it, methods become tools selected for a reason.

Methodology is an argument

Methodology is sometimes treated as the chapter where researchers describe what they did.

But a strong methodology does more.

It explains why the chosen approach is capable of answering the research question.

Suppose the question is:

How do doctoral researchers experience isolation during the thesis-writing stage?

A large questionnaire could tell us something useful.

It could estimate how commonly certain experiences occur.

But if we want to understand how researchers interpret isolation, how it develops, what situations intensify it and how individuals respond to it, detailed interviews may provide a different kind of evidence.

Neither method is inherently superior.

Their usefulness depends on the question.

This is why asking:

“Is quantitative research better than qualitative research?”

usually begins from the wrong premise.

The meaningful question is:

“Which approach produces the kind of evidence this research problem requires?”

Good methods cannot rescue a weak question

Imagine an impeccably administered survey.

The sample is large.

The questionnaire is reliable.

The statistical analysis is technically correct.

The tables are beautifully presented.

But the research question itself is trivial, confused or based on an unjustified assumption.

The study can be methodologically competent and intellectually weak at the same time.

Sophisticated analysis does not automatically produce significant research.

The same is true of qualitative work.

Fifty excellent interviews cannot rescue a study that never established what it was trying to understand.

Methodological rigour matters enormously.

But rigour has to be applied to a worthwhile intellectual problem.

The danger of choosing software before research design

Technology can intensify this problem.

Researchers sometimes begin with the tool:

SPSS

R

Python

NVivo

AMOS

SmartPLS

and then construct a study around what that software can do.

This reverses the logic of research.

Software performs analysis.

It does not decide which question deserves investigation.

It does not decide whether your variables represent the concepts you think they represent.

It does not determine whether your evidence supports the conclusion you want to make.

A sophisticated model applied to poorly conceptualised research remains poorly conceptualised research.

Statistical tests do not create meaning

Students frequently ask:

“Which test should I use?”

But a statistical test only makes sense after several things are already understood:

What kind of question are we asking?

What variables are involved?

How have they been measured?

What assumptions can reasonably be made about the data?

What comparison or relationship matters?

Only then does the choice of test become meaningful.

Consider the difference between:

Is there a relationship between study time and examination performance?

and:

Do students who receive a new teaching intervention perform differently from students who do not?

and:

Which factors predict examination performance?

All might involve numbers.

But they are not asking the same question.

The analysis should follow the logic of the research.

Research design is the bridge

Between the research question and individual methods sits something larger:

research design.

The design is the overall logic connecting:

the problem,

the question,

the evidence,

the participants or data sources,

the method of collection,

the method of analysis,

and the conclusions the researcher hopes to justify.

Think of methodology not as a collection of techniques, but as a chain of reasoning.

A strong study should allow the reader to see why each decision follows from the one before it.

A questionnaire is not a methodology

This distinction is especially important.

Students frequently write:

“The methodology used in this research is a questionnaire.”

A questionnaire is normally a data-collection instrument.

It does not tell us:

why quantitative evidence is appropriate,

why the selected population matters,

how the sample was determined,

how concepts became measurable variables,

how validity was considered,

how the data will be analysed,

or what limitations follow from those decisions.

Likewise:

“Interviews” are not an entire methodology.

“SPSS” is not a methodology.

“Thematic analysis” is not an entire research design.

These are components within a broader intellectual structure.

Understanding those distinctions is part of learning how research actually works.

The literature should change the question

Another reason research begins before methodology is that the researcher does not initially know enough.

A good literature review does more than provide citations for the proposal.

It changes the researcher’s understanding of the problem.

You may discover that the relationship you planned to investigate is already well established.

You may discover contradictory findings.

You may find that an apparently simple concept has been defined in several incompatible ways.

You may encounter a theory that changes how you interpret the entire problem.

You may realise that your first research question was asking the wrong thing.

That is not failure.

That is precisely what reading is supposed to do.

Research questions should be allowed to mature

Beginning researchers sometimes feel that changing the research question means something has gone wrong.

In reality, early questions should often change.

A researcher begins with limited understanding.

Then reads.

Thinks.

Discusses.

Observes.

Reads again.

Concepts sharpen.

Assumptions become visible.

Some questions disappear.

Better questions take their place.

The research question gradually becomes more precise because the researcher has become more knowledgeable about the problem.

A question written on the first day of a project should not be treated as sacred.

Methodological sophistication should come later

Once the intellectual foundations are clear, methodological sophistication becomes extremely valuable.

Now questions such as these matter:

How should the sample be constructed?

Which measures are valid?

Which interview approach will generate useful evidence?

What biases must be controlled?

Which statistical model is appropriate?

How will themes be developed?

How can competing explanations be tested?

How should uncertainty be reported?

At this stage, methodological decisions become genuinely powerful because they are serving a clearly defined purpose.

There is no universally “best” methodology

Researchers sometimes want an authority to tell them:

Use this method.

But methodology cannot usually be separated from context.

A method may be excellent for one question and completely inappropriate for another.

Experiments can provide powerful causal evidence under appropriate conditions.

Surveys can reveal distributions and relationships across large populations.

Interviews can uncover interpretation, meaning and experience.

Ethnography can reveal social practices that participants themselves may struggle to articulate directly.

Historical analysis can reconstruct developments that cannot be experimentally reproduced.

Secondary datasets can answer questions that would be impractical to investigate through new data collection.

Each method opens some possibilities while closing others.

Research design is therefore always partly about trade-offs.

The researcher must understand what the method cannot tell them

A strong methodology section should not merely defend the chosen approach.

It should understand its boundaries.

If the research is cross-sectional, what cannot be concluded about change over time?

If the evidence is self-reported, what biases may exist?

If participants come from one institution, how far can the conclusions reasonably travel?

If the research is qualitative, what kind of generalisation is actually being claimed?

If observational evidence shows association, can causation legitimately be inferred?

Research becomes trustworthy not when researchers pretend their methods are perfect, but when they understand what those methods permit them to claim.

The most important methodological question

After decades of teaching and supervising research, I believe one methodological question deserves to be asked repeatedly:

Why?

Why this population?

Why this sample?

Why this measure?

Why interviews?

Why this statistical test?

Why this theoretical framework?

Why this interpretation?

Why should the evidence support this conclusion?

If a researcher can answer those questions coherently, methodology begins to become more than a chapter.

It becomes the logic of the investigation.

Research is a chain of justified decisions

Ultimately, strong research can be thought of as a sequence.

Interpretation → Argument → Conclusion

Each stage constrains the next.

And each conclusion is only as strong as the reasoning that produced it.

Before you ask “Which method?”, ask something else

Methodology deserves careful attention.

But it should arrive at the right point in the research process.

Before asking:

Which research method should I choose?

ask:

What is the problem?

What do we already know?

What remains uncertain?

Why does that uncertainty matter?

What exactly am I trying to understand?

Once those questions have credible answers, methodology becomes easier to think about.

Because now we are no longer choosing a technique in search of a research problem.

We are choosing the most defensible way to investigate a problem that we understand.

And that is why research begins before methodology.