Writing the Methodology Section

The methodology section is where you prove your study was done right. It’s the most detail-oriented part of any research paper — and ironically, the section where AI is least helpful for generating content, because only you know what you actually did.

But AI is excellent at structuring, checking completeness, and improving clarity.

🔄 Quick Recall: In the previous lesson, you built a literature review using AI-powered discovery and synthesis. Your lit review established the research gap. Now you’ll describe exactly how you investigated it.

What the Methodology Section Must Include

Every methodology section answers four questions:

1. What did you do? (Research design)
2. Who did you study? (Participants/sample)
3. How did you do it? (Procedures and materials)
4. How did you analyze it? (Data analysis approach)

Where AI Actually Helps

Using AI to Structure Your Methodology

I'm writing the methodology section for a [type of study] paper.

Study details:
- Research design: [experimental/survey/qualitative/mixed methods]
- Participants: [who, how many, how recruited]
- Procedures: [what happened step by step]
- Measures/instruments: [what tools or scales you used]
- Data analysis: [planned statistical or qualitative approach]

Generate a methodology section outline with the standard subsections
for a [field name] paper targeting [journal name or style]. Flag any
elements I might be missing.

The AI will suggest subsections like: Research Design, Participants, Instruments, Procedures, Data Analysis, and Ethical Considerations. It may flag missing details — “You haven’t mentioned IRB approval” or “What was the response rate for your survey?”

✅ Quick Check: You’re writing a methodology section for a survey study. The AI suggests you include “response rate.” Why does this matter? (Answer: Response rate affects the generalizability of your results. A 15% response rate means your sample may not represent your target population. Reviewers will flag a missing response rate as a methodology weakness.)

Writing Each Subsection

Research Design

State your approach clearly and justify it:

Help me write a concise research design paragraph:
- Study type: [randomized controlled trial / quasi-experimental /
  cross-sectional survey / qualitative interviews / etc.]
- Why this design: [why it's appropriate for your research question]
- Key variables: [independent, dependent, control]

AI can help you articulate why your design fits your question, but verify the justification is accurate for your field.

Participants

Be specific:

Write a participants section covering:
- Total sample: [N]
- Demographics: [age range, gender distribution, relevant characteristics]
- Recruitment: [how participants were found]
- Inclusion/exclusion criteria: [who qualified, who didn't]
- Consent process: [how informed consent was obtained]
- Attrition: [how many dropped out and why, if applicable]

Procedures

Describe what happened in chronological order. The replicability test: could someone follow these steps and repeat your study?

I need to describe my research procedures clearly enough for replication.
Here's what happened in order:
[List your steps in bullet points]

Rewrite as flowing prose in past tense, academic style. Flag any steps
where more detail is needed for replicability.

Data Analysis

This is where AI can suggest statistical approaches — but always verify:

My research design:
- [Describe your study briefly]
- Independent variable(s): [type and levels]
- Dependent variable(s): [type and measurement scale]
- Sample size: [N]

Suggest appropriate statistical tests and justify each choice.
Include assumptions I need to verify for each test.

✅ Quick Check: AI suggests a paired t-test for your pre-post study design. What assumptions do you need to verify before using it? (Answer: The differences between paired observations should be approximately normally distributed, the data should be measured on an interval or ratio scale, and the pairs should be independent of each other. With small samples (n < 30), the normality assumption is especially important.)

AI as Methodology Reviewer

After writing your methodology, use AI as a reviewer:

Review this methodology section for completeness and clarity:

[Paste your methodology]

Check for:
1. Missing standard elements (design justification, sample size rationale,
   ethical approval, limitations)
2. Ambiguous descriptions that would prevent replication
3. Inconsistencies between stated design and described procedures
4. Appropriate statistical test selection for the stated design

This catches common oversights: forgetting to mention ethical approval, not stating how missing data was handled, or describing procedures out of order.

Common Methodology Mistakes

Practice Exercise

1. Write a bullet-point list of your study’s key details (design, participants, procedures, analysis)
2. Use AI to generate a methodology section outline
3. Draft each subsection, using AI for structure and clarity — not content invention
4. Run the completed section through AI as a “methodology reviewer”
5. Verify all statistical test suggestions against your data’s actual properties

Key Takeaways

• The methodology section describes what you did — AI can structure and improve clarity, but content comes from you
• The replicability test: another researcher could repeat your study from your description
• AI is useful for suggesting statistical tests, but always verify your data meets the assumptions
• Use AI as a reviewer to catch missing elements (ethical approval, sample size rationale, limitations)
• Common oversights: no power analysis, vague procedures, wrong statistical test

Up Next

In the next lesson, you’ll draft the Results and Discussion sections — presenting your findings and interpreting what they mean in the context of existing literature.