How to Write a Dissertation Methodology Chapter for Quantitative Research (After 12 Years of Doing This With Students)

How to write a dissertation methodology chapter for quantitative research is the question I get asked more than any other, usually at 11 pm, two weeks before submission. Over twelve years of guiding dissertation students across the UK, the US and back home in India, I have read the same guides you have. Most of them, including well-known names like GradCoach, tell you to “take inspiration from other studies’ methodological approaches.” I disagree with that advice on its own.

Matching a precedent tells your examiner you copied a structure. It does not tell them you understood why that structure fits your research question. If you are looking for how to write quantitative research methodology chapter guidance you can actually defend in a viva, not just a template to copy, that is exactly what follows.

What Is a Methodology Chapter (And Why Examiners Obsess Over It)

A dissertation methodology chapter is the section where you explain, justify and defend every decision you made to collect and analyse your data. Not describe. Justify. That one word is where most students lose marks.

Examiners do not read this chapter to find out what you did. They read it to check whether your results can be trusted. If your sampling was biased or your test choice was wrong, everything in your results chapter becomes questionable, no matter how well it is written.

Methodology vs Methods: The Difference Nobody Explains Properly

Methodology is your overall approach and the reasoning behind it, positivist, deductive, quantitative. Methods are the specific tools you used, a survey, SPSS, a t-test. Most guides use these words interchangeably. That is sloppy, and examiners notice.

Where It Sits in Your Dissertation

It comes right after your literature review and before your results chapter. Your literature review should have already hinted at the gap your methodology is designed to fill. If it has not, go back and fix that link first, because your committee will ask about it in the viva.

How to Write a Dissertation Methodology Chapter for Quantitative Research: What to Include

Every dissertation methodology chapter, quantitative research or otherwise, needs these seven parts, in this order:

1. Research design and philosophy
2. Population and sampling strategy
3. Data collection instrument
4. Data analysis plan
5. Validity and reliability
6. Ethical considerations
7. Limitations

Skip one, and your committee will find it. I have seen a strong dissertation get sent back purely because the student forgot to name the sampling technique, even though the sample size was mentioned. Naming things matters here.

Length: 2,500 to 4,000 words for most master’s dissertations. If you are wondering how to write a methodology chapter for PhD quantitative research specifically, expect 5,000 to 8,000 words, because a PhD chapter also defends your epistemology in its own right and usually separates validity and reliability into fuller, standalone discussions rather than one shared section.

Tense: past tense throughout, because you are describing what you did, not what you are doing.

Step 1: State Your Research Design and Philosophy

Start by naming your paradigm. Most quantitative dissertations sit in positivism, using deductive reasoning to test a hypothesis drawn from existing theory. Say that in one sentence, then justify it against your research question, not against what your supervisor prefers. This basic pairing of philosophy and design is central to Creswell’s Research Design: Qualitative, Quantitative, and Mixed Methods Approaches, and it still holds up as the standard reference most supervisors expect you to know.

Quantitative vs Qualitative, and Positivism vs Interpretivism

If you are still deciding your overall approach, the short version is this. Quantitative work tests a hypothesis using numbers and statistics, and usually sits within positivism, which assumes reality can be measured objectively. Qualitative work explores meaning and experience through words, and usually sits within interpretivism, which assumes reality is shaped by perspective. Once you have picked quantitative, you have already picked positivism as your default philosophy in most disciplines.

If your study is measuring change over time versus a single snapshot, this is also where you decide your design type. I still see students confuse these two designs in their proposal defence, so check how a cross-sectional design differs from a longitudinal one before you commit to either.

Step 2: Define Your Population and Sampling Strategy

This is where most rejections happen. You need three things: who your population is, how you selected your sample from it, and why that sample size is defensible.

Probability vs Non-Probability Sampling: Which to Choose

Probability sampling, random, stratified, systematic, gives you generalisable results but needs a proper sampling frame listing your whole population. Non-probability sampling, convenience, purposive, is faster and more realistic for student timelines, but you must admit its limits rather than hide them. Pick probability sampling when generalisability is central to your research question. Pick non-probability sampling when your population is hard to fully list or access, and say so plainly in the chapter rather than skating past it.

A Worked Example on Sample Size

A student of mine researching employee engagement in Indian IT firms wanted 500 responses “because it sounded rigorous.” We ran a G*Power calculation instead, using a medium expected effect size, following the conventions in Cohen’s 1988 Statistical Power Analysis for the Behavioral Sciences, and a 0.05 significance level, and landed at 138 as the minimum for her regression model. She collected 160. Her methodology chapter now had a number backed by a formula, not a guess, and her examiner did not question it once.

If your data is coming from records rather than surveys, your sampling logic changes again. Get this right before you write the section on choosing between primary and secondary data for your dissertation.

Step 3: Describe Your Data Collection Instrument

State exactly what you used, a questionnaire, a validated scale, an experiment, and where it came from. If you adapted a scale from an existing study, cite it in full APA style, author, year and original scale name, rather than a passing mention. If you built one from scratch, say so honestly. Examiners are more forgiving of a self-developed instrument than a hidden one.

Validating Your Instrument Before You Use It

Mention your pilot study here. Even ten respondents used to test clarity of wording counts, and leaving it out makes your instrument look untested. Name the platform you administered it on too, Qualtrics and Google Forms are the two examiners see most often, since this affects how you describe data storage in your ethics section. If you are unsure whether to adopt an existing scale or build your own, this comparison of validated scales against custom instruments will save you a rewrite later.

Step 4: Write Your Data Analysis Plan

This is the section students fear the most, and it is the shortest to write once you know the logic. Match your research question type to a test, not the other way round.

Which Statistical Test Should You Use? A Quick Decision Table

Your research question is about	Likely test
Difference between two groups	Independent t-test
Difference across three or more groups	ANOVA
Relationship between two variables	Pearson or Spearman correlation, depending on your data type
Predicting one variable from others	Regression, after checking multicollinearity with VIF
Pre-post scores on the same group	Paired t-test or repeated measures ANOVA

Before you pick a test, confirm your data actually meets its assumptions. Skipping this step is the single most common flaw I find in draft chapters. Run a Shapiro-Wilk test for normality first, every time, even when you are confident the data looks fine on a histogram.

SPSS vs R vs Python: What Actually Matters in Your Write-Up

I disagree with guides that say the software choice does not matter for this chapter. It does. SPSS is point-and-click and easier for a first dissertation, but it hides some of the assumption checks behind menus you have to know to open. R is free, has a far wider range of tests, and forces you to see every assumption check explicitly, at the cost of a steeper learning curve.

Python is worth it only if your analysis overlaps with machine learning or a very large dataset, which is rare in a standard quantitative dissertation. Whichever you pick, name it along with the version, because an examiner checking replicability wants to know exactly what you ran it in. If SPSS is your tool, a structured SPSS tutor guide is worth working through before you touch your real dataset.

Step 5: Address Validity, Reliability and Ethics

How Reliable Is Reliable Enough?

For quantitative work, reliability usually means internal consistency, measured by Cronbach’s alpha. The widely cited threshold, from Nunnally’s 1978 Psychometric Theory, is 0.7 as acceptable and 0.8 as good. Run the number yourself rather than assuming it, using a Cronbach’s alpha calculator, and report the actual value, not just the threshold.

A nursing student I worked with built a custom 6-item scale for measuring burnout and assumed it would be fine because it “made sense.” Her alpha came back at 0.58. We cut two poorly worded items and reran it, landing at 0.76. Without that check, her results chapter would have rested on a scale that could not be trusted, and no examiner would have missed it.

What Your Ethics Statement Must Include

State how you got informed consent, how you protected anonymity, and whether your university’s ethics board approved the study. Even desk-based quantitative research using secondary data needs a line confirming you had permission to use that data source, since “it’s not primary data” is not an ethics exemption on its own.

Step 6: State Limitations Without Undermining Your Study

Every study has limits. A small sample, a single-country context, self-reported data. Name two or three honestly, then explain why they do not invalidate your findings. Students who skip this section entirely look less credible than students who admit a flaw and defend around it.

Common Mistakes That Get Chapters Sent Back

From twelve years of reviews, these repeat constantly, and most of them resurface again in the viva if they slip through at submission:

• Describing a method without justifying it against the research question
• Confusing sampling technique with sample size
• Choosing a statistical test before checking its assumptions
• Copying a generic ethics paragraph that does not match the actual study
• Writing the chapter in present tense

Can You Use AI to Draft Your Methodology Chapter?

Students ask me this constantly now, and I will be direct. Using AI to structure your thinking or check your logic is fine. Using it to generate your justifications wholesale is not, because AI tools regularly fabricate sample size formulas and cite tests that do not fit your data, and examiners are now trained to spot the pattern.

If your supervisor has already flagged your writing as AI-sounding, read this before you touch the chapter again: what to do if you are accused of using AI in your dissertation.

FAQ

Getting a second pair of eyes on your test selection, with the analysis, interpretation and final decisions staying yours, is exactly what most university guidelines expect and welcome. If you would rather have someone sanity-check your test selection or actually run the analysis with you before your supervisor sees it, that is exactly what our SPSS tutoring sessions are built for.

About the Author: Written by a dissertation and statistics expert with over 12 years of research guidance experience across R, Python, SPSS and Stata. Connect on LinkedIn.