How to Write a Methodology Chapter for a Quantitative Study 2026

The short answer: document every design decision in enough detail that another researcher could repeat your study from your chapter alone. A quantitative methodology chapter is not a list of methods, it is the argument that each method was the right one for the research question it serves, backed by sample size, instrument and ethics evidence another researcher could check.

Most students lose marks in one place: they write what they did without justifying why they did it. Fix that and the rest is assembly. The exact chapter number shifts by institution, Chapter 3 in a five-chapter dissertation and Chapter 4 in a seven-chapter structure, so check your department handbook before you worry about formatting.

Table of Contents

1. 1What You Need
2. 2Step-by-Step: How to Write a Methodology Chapter for a Quantitative Study
3. 3Step 1: Explain the research design
4. 4Step 2: Define the study setting and population
5. 5Step 3: Describe the sampling strategy
6. 6Step 4: Explain data-collection instruments and procedures
7. 7Step 5: Present the quantitative data-analysis plan
8. 8Step 6: Address ethics and quality assurance
9. 9Step 7: Write and edit the chapter
10. 10Common Mistakes
11. 11Frequently Asked Questions
12. 12Is the methodology chapter 3 or 4?
13. 13Should the methodology chapter be in past tense or present tense?
14. 14How long should a methodology chapter be?
15. 15Do I report Cronbach’s alpha in the methodology chapter or the results?
16. 16How do I justify a sample size when no population data exists?
17. 17What is the difference between methodology and methods in a chapter?
18. 18Conclusion

What You Need

You cannot write this chapter from memory. You need the documentation trail first, and most of it already exists in some form by the time you sit down to draft.

• The approved proposal or research protocol. This is your baseline. Every research question and objective in the chapter should trace back to one line in the proposal.
• Ethics documentation. Approval reference number, approving body, and the date. If you are still awaiting approval, say so and describe the process.
• The instrument itself. The full questionnaire with item numbers, or the scale name, author, number of items and scoring.
• Sampling materials. The sampling frame, the recruitment message, and a record of how many people were contacted.
• The analysis plan. A mapping of each research question to a variable set and a test, plus the significance level you committed to.
• Software details. Package and version, plus any add-ons. IBM SPSS Statistics 29.0 and R 4.3.1 with the lme4 package are the level of specificity that reads as credible.
• Supervisor and university requirements. Page limits, mandatory headings, referencing style and submission format.

A one-page design summary is the fastest way to spot whether you are ready. If you cannot fill that page, you are not ready to draft, and no amount of editing will fix a chapter with nothing behind it.

Step-by-Step: How to Write a Methodology Chapter for a Quantitative Study

Step 1: Explain the research design

Name the design in the first two sentences of the chapter, then justify it. A cross-sectional correlational survey design was selected because the study tests whether job satisfaction predicts turnover intention among registered nurses, and no manipulation of variables is possible in a single organisation.

Quantitative designs separate into a few families, and the justification is what earns credit, not the label.

• Experimental. Researcher manipulates a treatment and assigns participants to conditions. Justify it when you can control who receives what and randomise.
• Quasi-experimental. Manipulation without full randomisation, such as intact existing groups. Expect a section on threats to internal validity.
• Correlational. Measures variables without manipulation. The right default for relationship questions.
• Comparative. Compares two or more groups, for example pre-service and in-service teachers.
• Longitudinal versus cross-sectional. Follows the same participants over time, or takes one measurement. Say which you chose and why.
• Descriptive and exploratory. Estimates prevalence or averages, or tests a new topic with few prior assumptions.

The review of the design should be one or two paragraphs, not a survey of every design in the literature. Link the design to the research questions directly: question, design element that answers it, done.

Philosophical commitment usually gets one honest paragraph. Positivism fits quantitative work that assumes measurable reality and replicable results. Pragmatism fits when you mix quantitative outcomes with process evidence, and most applied studies sit here in practice even if they never name it.

Step 2: Define the study setting and population

Describe where and when the study ran, then define the population precisely enough that someone else could identify the same group. Naming the target population as all nurses working in direct patient care in the region, not nurses generally, is what makes the sampling section defensible.

Three populations get confused constantly. Separate them in your writing.

Population type	Definition	Example
Target population	The full group your findings are meant to describe	All registered nurses in direct patient care across the three participating hospitals
Accessible population	The part of the target population you can actually reach	Permanent and bank nurses on the staff lists at the three hospitals
Sample	The participants actually recruited and measured	287 nurses who returned a complete questionnaire

State inclusion and exclusion criteria as explicit rules with reasons. Adults aged 18 or over, employed at least three months, not on extended leave during the collection window. Say why each exclusion exists, because exclusions that look arbitrary invite questions you would rather not field at viva.

The setting paragraph also carries the dates, the mode of data collection, and any conditions that could have shaped responses, such as surveys distributed at shift handover or during a staffing shortage.

Step 3: Describe the sampling strategy

Pick the sampling method, name the frame, show how participants were selected, and justify the sample size. These are the four things examiners look for, and the third one is where most chapters go quiet.

Method	How selection happens	Use it when	Main limitation
Simple random	Every member has an equal known chance	A complete, accessible list exists	Impractical for large or mobile populations
Stratified	Divided into strata, then sampled within each	Subgroups must be represented proportionally or equally	Needs accurate knowledge of strata sizes
Cluster	Whole groups are sampled, all members taken	No usable list, but definable natural groups	Design effect raises the sample needed
Systematic	Every kth case from a list	Large ordered lists, low cost	Periodicity in the list can bias the sample
Convenience	Whoever is easiest to reach	Exploratory work or pilot studies	Unknown bias, weak generalisation
Purposive	Deliberately chosen to fit criteria	Hard-to-reach populations, qualitative strands	Selection judgement is hard to defend statistically
Snowball	Participants recruit further participants	No roster exists for the population	Over-represents well-connected members

Justify sample size with a power analysis, not with a round number. In G*Power for a two-sided independent-samples t-test, typical settings are a significance level of 0.05, power of 0.80, and a medium effect size of d = 0.5, which returns a required sample of 64 per group, or 128 in total. Add 15 to 20 percent for non-response and state the final target and the number actually achieved.

The hard case is a population nobody has counted. A student researching small business owners who use e-bikes for delivery can often get total small-business counts but has no data on e-bike use, which leaves a power analysis with no effect-size basis from prior literature. The workable answers are to state the assumption explicitly and defend it, to triangulate from adjacent literature on similar adoption behaviours, or to treat the study as an exploration and report the achieved sample with its achieved power. In a mixed-methods study, saturation from the qualitative strand can justify the quantitative target, and it should be written as a deliberate design decision rather than an excuse.

Report the response rate as invited divided by returns. A 287 of 412 invitation giving a 69.7 percent response rate reads as honest and lets a reader judge non-response bias themselves.

Step 4: Explain data-collection instruments and procedures

For each instrument, give the name, the author or scale, the number of items, the response format, the scoring range, and the source of permission. If you translated or adapted a scale, report the forward and back-translation procedure and who checked the content.

Report the evidence you have for measurement quality, in this order of preference.

• Content validity. Expert review by a panel, ideally 3 to 5 people, stating who they were.
• Construct validity. Factor analysis results, or convergent and discriminant evidence from established literature.
• Criterion validity. Comparison against a known outcome or external measure.
• Reliability. Cronbach’s alpha for each scale, plus pilot test-retest figures where time allows.

Cronbach’s alpha belongs in the methodology chapter when the pilot produced it, and again in the results when the full dataset produces it. Say which is which. A value of 0.91 for the four-item burnout scale from the pilot, and 0.89 across the full sample, is exactly the detail examiners want, and it separates the two chapters cleanly.

A pilot study is the cheapest credibility available. Run 30 to 50 cases with the real instrument and real conditions, and report what changed as a result. Where a question was reworded after a supervisor comment, that is a strength, not an embarrassment.

Describe administration in reproducible detail: mode, platform, time taken, whether the researcher was present, reminders, and what participants were told about anonymity. Note the contact and response rate for online instruments too, since both bear on coverage bias.

Step 5: Present the quantitative data-analysis plan

Map every research question to a variable set and a test, and show the mapping in a table. This single table does more work than any other part of the chapter because it demonstrates that nothing in the analysis was decided after seeing the data.

Research question	Variable types	Test	Assumption checks	Software
Does burnout predict turnover intention?	Continuous predictor, continuous outcome	Multiple linear regression	Linearity, homoscedasticity, VIF under 5, normality of residuals	SPSS 29.0, regression
Is there a difference in satisfaction between two nursing groups?	Categorical grouping, continuous outcome	Independent-samples t-test	Shapiro-Wilk for normality, Levene’s test for equal variances	SPSS 29.0, t-test
Does shift type relate to reported stress level?	Categorical, ordinal outcome	Chi-square test of independence	Expected counts of 5 or more per cell	R 4.3.1, chisq.test
Which factors predict intention to leave?	Mixed predictor types, binary outcome	Logistic regression	Linearity in logit, no extreme outliers, events per variable of 10 or more	Stata 18, logit

Set the significance level before reporting results and keep it, normally 0.05. Say what you will do about missing data, for example listwise deletion below 5 percent missing, mean substitution otherwise, or multiple imputation if missingness is non-random. Outliers get a stated rule, such as removal beyond three standard deviations, applied only to continuous variables and justified rather than chosen to tidy a histogram.

The methods chapter declares which assumption checks were planned. The results chapter reports their outcomes. Writing normality tests in the methods chapter is a mistake; the method is testing for normality, the result is what the test returned.

Cover data preparation too: coding of open items, recoding of reverse-scored items, handling of straight-lining and duplicate submissions, and the transformation of skewed variables. Then state the software, the version, and any package used, because someone trying to reproduce your work needs the exact build.

Step 6: Address ethics and quality assurance

Report ethical approval in a form that can be verified: the approving body, the reference number, and the approval date. Say how consent was obtained and what participants were told about their right to withdraw before analysis began.

Cover confidentiality and anonymity as separate claims, because they are not the same. Confidentiality limits disclosure, anonymity means no identifier is retained at all. If quotes are impossible in a quantitative design, do not include a confidentiality paragraph you cannot support.

Include how the data itself is protected: encrypted storage, restricted access to identifiable files, a stated retention period, and the destruction procedure afterwards. Data protection impact assessments matter where you handle special category data such as health records.

On quality, explain what you did to reduce bias rather than asserting that none existed. Randomisation where possible, blinding of assessors, a neutral recruitment message, an anonymous response setting, and a pre-registered analysis plan all belong here. Then state the known limitations honestly: non-response bias, self-report bias, social desirability, and the risk that convenience sampling restricts generalisation.

Students routinely pick instruments because they are easy to find rather than because they fit the construct, and reviewers have flagged exactly that pattern in submitted theses. Choose on validity evidence, and say why you chose this one over the alternative.

Step 7: Write and edit the chapter

Write completed procedures in the past tense: a survey instrument was distributed, not a survey instrument is distributed. Future or present tense belongs in a proposal that describes planned work, which is the one legitimate reason to switch. Pick one convention for the whole chapter and hold it.

Apply the replication test line by line. For each paragraph, ask whether a researcher with access to your raw data could repeat the procedure, rerun the analysis and get the same numbers. Anything that fails that test either gains detail or moves to the results chapter.

Align every sentence with your research questions. If a paragraph does not serve a question, a research objective, or a defensible methodological choice, cut it. Supervisors reject chapters that read as a menu of methods rather than an argument, and length is not what they are checking.

On length, expect roughly 15 to 25 pages for a doctoral chapter, 10 to 15 for a master’s thesis, and 3 to 5 for an undergraduate project, though many institutions set a hard page limit and that limit wins. Detail should match how standard the method is in your discipline: an ordinary cross-sectional survey needs a paragraph, a bespoke adapted scale or a custom instrument needs several.

Reference in APA 7th edition, citing the originators of the design, the scale and the test rather than secondary summaries. Finally, proofread for the things word counters miss: every table and figure referenced in text before it appears, abbreviations defined at first use, and one tense throughout.

Common Mistakes

Most lost marks come from the same eight problems. Each has a specific fix.

1. Listing methods without justifying them. Fix: after every method, add a sentence connecting it to a named research question. If you cannot, the method does not belong.
2. Vague design descriptions. Fix: name the design, the setting, the dates and the frame. Replace a generic “a survey was conducted” with what was asked, how, to whom and when.
3. Sample size justified by a round number. Fix: run the power analysis, report the inputs and the output, then add a non-response allowance. If no effect size exists, state the assumption openly and defend it.
4. Unsupported validity claims. Fix: attach evidence. Content validity means an expert panel and you name their expertise. Reliability means a Cronbach’s alpha per scale from a named pilot sample.
5. Unexplained statistical choices. Fix: publish the mapping table. One row per question, with the test, the assumption checks and the software command.
6. Mixing methods with results. Fix: methods describe what you will do, results report what happened. This is the single clearest line examiners use to divide the two chapters.
7. Ethics treated as a formality. Fix: give the approval body, reference number and date, and describe consent, anonymity and data protection in specific rather than general terms.
8. Tense drift mid-chapter. Fix: run one search for verbs ending in -ed and -ing within your methods sections, then normalise to past tense unless the study is prospective.

Belongs in the methodology chapter	Belongs in the results chapter
The decision to use Cronbach’s alpha	The value of Cronbach’s alpha
Planned tests for assumption violations	The outcome of those tests
Response rate calculated as invited versus returns	Demographic and descriptive statistics of the sample
Software and version used	The coefficient, p-value and confidence interval
Stated limitations and how they were addressed	Whether the limitations actually affected findings

Before you submit, run this checklist: every research question appears with a matching analysis; the sample size has a stated basis; each instrument has validity or reliability evidence; the ethics reference is present; every table is referenced in text; and the tense is consistent. If any answer is no, fix it now rather than in a viva.

Frequently Asked Questions

Conclusion

A quantitative methodology chapter holds together through one repeated move: name a design decision, then say which research question it serves and what evidence backs it. Design, population, sampling with a power-based sample size, instruments with validity and reliability figures, a published mapping from question to statistical test, and verified ethics approval give a reader everything they need to trust the results that follow.

Start with the one-page design summary. Write the research questions on the left, the design, sample, instrument and test for each on the right, and every gap shows up immediately. Once that page is complete, drafting becomes transcription, and the chapter holds together on its own.