How to Write Chapter 3 of a PhD Thesis Proposal (Research Methodology): The Complete Guide

Chapter 3 is where your thesis proposal either earns your committee’s confidence, or loses it. This is the chapter where examiners stop asking “what are you studying?” and start asking “how, exactly, will you find out and can I trust the answer?”

The format of a PhD or Masters thesis proposal varies slightly from one institution to another, but in the vast majority of programs especially the standard 3-chapter structure used across most African universities and many international ones, Chapter 3 is the research methodology chapter.

This guide walks through every major section of Chapter 3, in the order examiners expect to see them, with practical examples throughout.

What Chapter 3 Actually Needs to Do

Before the section-by-section breakdown, it helps to understand the one job Chapter 3 has: prove that your methods can actually answer the research questions you posed in Chapter 1.

Examiners read Chapter 3 asking three questions:

1. Does this design fit the objectives?
2. Can this sample size and sampling approach support the claims the student wants to make?
3. Is there a credible plan for collecting and analysing the data?

Every section below exists to answer one of those three questions. Keep that in mind and the chapter becomes much easier to structure; you’re not filling in a template, you’re building a case.

Research Design: Quantitative, Qualitative, or Mixed-Methods?

This section states and justifies the overall research design the study will use. The choice should be driven directly by your research objectives and research questions from Chapter 1, not by personal preference or convenience.

Quantitative research design

The study collects, analyses, and presents numerical data in the form of statistics: descriptive, inferential, or both. Quantitative designs work well when your objectives ask “how much,” “how many,” or “is there a relationship/effect between X and Y.”

Qualitative research design

The study collects, analyses, and presents non-numerical data: words, opinions, narratives, and lived experiences. Qualitative designs suit objectives that ask “why” or “how” a phenomenon is experienced or understood.

Mixed-methods research design

The study combines quantitative and qualitative approaches, typically because some research questions call for numbers and others call for narrative depth. Mixed-methods studies should specify whether they are concurrent (collected at the same time) or sequential (one phase informs the next), and which strand; quantitative or qualitative takes priority.

Each design has trade-offs in generalisability, depth, time, and resources required. Whatever you choose, state explicitly why it fits your specific objectives. This single sentence of justification is one examiners specifically look for and one many students skip.

Population and Sampling

Population of study refers to the entire set of subjects relevant to your research problem. If the population is small, you may include all subjects in the study (a census). If it’s large, including everyone becomes impractical in terms of time and resources — so you draw a sample.

A sample is a sub-set of the population from which data will be collected to draw conclusions about the population as a whole.

Example: A study investigating the effects of the COVID-19 pandemic on micro and small enterprises (MSEs) in Kenya has a population consisting of every MSE in Kenya which includes thousands of businesses spread across the country. Collecting data from all of them is not feasible, so the researcher draws a sample instead. The required sample size will depend on whether the study is quantitative, qualitative, or mixed-methods: quantitative studies typically need larger samples to support statistical generalisation, while qualitative studies prioritise depth over size.

Sampling is the process by which a sample is drawn from a population, and it falls into two broad categories: probability (random) and non-probability (non-random) sampling, which are covered in detail below.

How to Determine Your Sample Size

This is one of the most common gaps in student proposals and one of the fastest ways to get Chapter 3 sent back. “I will sample 100 respondents” without justification is not a methodology decision; it’s a guess. Examiners want to see the formula or table you used and the assumptions behind it.

For quantitative studies with a known population

Two formulas are widely used and accepted across most institutions:

Slovin’s Formula

n = N / (1 + N·e²)

Where n = sample size, N = population size, and e = margin of error (commonly 0.05 for a 95% confidence level).

Cochran’s Formula (useful for large or unknown populations)

n₀ = (Z²·p·q) / e²

Where Z is the standard score for your desired confidence level (1.96 for 95%), p is the estimated proportion of the population with the characteristic being studied (use 0.5 if unknown, since this maximises the sample size), and q = 1 − p.

Many students also reference the Krejcie and Morgan (1970) table, which gives a recommended sample size directly from a known population size without requiring the formula to be calculated manually. This is useful when your institution accepts published sampling tables as a citation.

For qualitative studies

Sample size is not calculated with a formula. Instead, justify your sample using the concept of data saturation: the point at which additional interviews or focus groups stop producing new themes or information. State an anticipated range (for example, “12–15 in-depth interviews, continuing until saturation is reached”) rather than a single rigid number.

What examiners actually check

Whichever approach you use, Chapter 3 should show your working: the formula, the values you plugged in, and the resulting number, not just the final sample size in isolation.

Sampling Techniques: Probability vs Non-Probability

There are two broad categories of sampling techniques. Which one you use depends on your research design.

Probability (random) sampling

In random sampling, every subject in the population has a known, equal chance of being selected. Results from the sample can be generalised to the wider population, especially with a sufficiently large sample size. Random sampling is used primarily in quantitative studies, and includes:

• Simple random sampling: every subject has an equal, independent chance of selection (e.g. drawing names from a list using a random number generator).
• Systematic sampling: subjects are selected at a fixed interval from a list (e.g. every 10th name on a register).
• Stratified sampling: the population is divided into relevant sub-groups (strata) such as gender, region, or income bracket and a random sample is drawn from each stratum, ensuring proportional representation.
• Cluster sampling: the population is divided into naturally occurring clusters (e.g. villages or schools), and entire clusters are randomly selected rather than individuals.

Non-probability (non-random) sampling

In non-random sampling, subjects are selected deliberately rather than randomly, so they do not have an equal chance of selection. It’s also called purposive sampling, since the sample is chosen for a specific purpose. Non-random sampling is used primarily in qualitative studies, and includes:

• Purposive sampling: subjects are selected because they have specific knowledge or characteristics relevant to the research question.
• Convenience sampling: subjects are selected based on ease of access (fastest to arrange, but weakest for generalisability).
• Snowball sampling: existing participants refer the researcher to other potential participants; useful for hard-to-reach populations.
• Quota sampling: the researcher sets a target number of subjects for specific sub-groups, then samples non-randomly until each quota is filled.

State clearly which technique you are using, and why it fits your design. A stratified sample makes sense for a study needing proportional regional representation; purposive sampling makes sense for a qualitative study needing information-rich key informants.

Data Collection Methods and Tools

This section discusses in detail the type of data you’ll collect: primary, secondary, or both, and exactly how you’ll collect it from your sample. The methods and tools depend on your research design.

Questionnaires

Questionnaires are mostly used to collect quantitative data. They are structured, using closed-ended questions. Four common question types appear in questionnaires:

• Numerical questions: e.g. “How many children do you have?”
• Two-option (dichotomous) questions: e.g. “Does your household have a radio? 1. Yes 2. No”
• Multiple choice questions: e.g. “What is your highest level of education? 1. No education 2. Primary 3. Secondary 4. Tertiary”
• Likert scale questions: e.g. “Rate your satisfaction with the water services board: 1. Very dissatisfied … 5. Very satisfied”

Questionnaires can be facilitated (the researcher administers it, face-to-face or by phone) or self-administered (the respondent completes it independently, delivered by hand, post, or email). Each delivery mode carries trade-offs in response rate, cost, and data quality.

Interviews

Interviews are oral discussions between researcher and respondent. Unlike questionnaires, interviews are semi-structured: the researcher uses an interview guide with a set of core questions, but subsequent questions and discussion flow are shaped by the respondent’s answers, meaning the interview’s path varies from person to person.

Focus group discussions (FGDs)

Where interviews are one-on-one, FGDs are held with a group of respondents relevant to the problem under investigation. Participants should be selected carefully to represent diverse perspectives.

Example: In the MSEs-and-COVID-19 study, a focus group might include a female-owned enterprise, a male-owned enterprise, a youth-owned enterprise, a family-run enterprise, a non-family enterprise, customers, and an official from the Micro and Small Enterprises Authority (MSEA) producing richer discussion than any single perspective alone.

Observation

Observation comes in two forms. In participant observation, the researcher immerses themselves in the study environment (e.g. working inside one of the MSEs to observe daily operations first-hand). In non-participant observation, the researcher remains outside the environment, observing from a distance (e.g. tracking how many customers visit a business per day without direct involvement). An observation checklist should guide what is observed and how often.

Document review

Here, the student obtains and analyses documents relevant to the study. For the MSEs example, this might include the MSE Policy of Kenya or the Strategic Plan of the Micro and Small Enterprises Authority. Document review is useful for understanding the current state of affairs surrounding the problem under investigation.

Validity and Reliability (Don’t Skip This)

A section many students leave out entirely and one examiners increasingly ask about directly during defense. Without it, your data collection tools look untested.

Validity

Validity asks: does the instrument measure what it claims to measure? Common types to address:

• Content validity: does the tool cover the full breadth of the concept being measured? Often established by having supervisors or subject experts review the instrument.
• Construct validity: does the tool truly capture the underlying theoretical concept?
• Face validity: does the tool appear, on the surface, to measure what it’s supposed to?

Reliability

Reliability asks: would this instrument produce consistent results if repeated? For quantitative instruments, this is often demonstrated using Cronbach’s Alpha (a coefficient above 0.7 is generally considered acceptable) calculated from a pilot test.

Pilot testing

Most institutions expect a brief mention of a pilot study, which entails testing your questionnaire or interview guide on a small group similar to (but not part of) your actual sample, to catch ambiguous questions, timing issues, or translation problems before full data collection begins.

Ethical Considerations

This section highlights the ethical safeguards that will be followed during data collection. This is especially important for studies involving human subjects. Ethical considerations vary by study but commonly include:

• Informed consent: the researcher must obtain consent before data collection begins, explaining what the study is about, how the respondent was selected, and what benefits (if any) exist, before requesting permission to proceed. Consent may be written or oral.
• Compensation for participation: participation should be voluntary, though some studies offer monetary compensation. Respondents should be informed of any compensation plans, but only after they have participated — not before, to avoid coercion.
• Confidentiality: respondents should be assured their responses will remain confidential.
• Dissemination of findings: there should be a plan to share results with participants, e.g. through validation workshops or written summaries.

Most academic institutions also require students to obtain formal ethical clearance from an institutional review board or ethics committee before data collection. Check your institution’s specific requirement and timeline early because clearance can take weeks.

Data Analysis

This section explains how the collected data will be analysed. Methods vary depending on whether the data is quantitative or qualitative.

Quantitative data analysis

The focus is on numbers, analysed through descriptive and inferential statistics.

Descriptive statistics is usually the first analytical step, covering:

• Measures of frequency (frequency tables, cross-tabulations)
• Measures of central tendency (mean, median, mode)
• Measures of variability (range, standard deviation, variance)

Inferential statistics goes further, testing whether sample results can be generalised to the wider population, or whether an intervention had measurable impact. Common techniques include:

• Tests for differences between groups: t-test, ANOVA, Chi-square test
• Tests for correlation or causation between variables: linear regression, logistic regression (logit, probit, multinomial logit/probit models)

Your choice of technique should be guided by your data type (a continuous dependent variable needs a different technique from a categorical one) and, above all, by your research questions. The analysis must be capable of actually answering the questions you posed.

Qualitative data analysis

Analysis involves working through the content of interviews and FGDs such as audio recordings and hand-written notes. Recordings should be transcribed and notes organised before analysis begins. The process typically involves coding the data, indexing it, and framing it to identify emerging themes.

Software

State which software you’ll use for analysis. Common choices include SPSS and Stata for quantitative data, and NVivo for qualitative data.

Limitations of the Study

The final section of Chapter 3 discusses the study’s potential limitations and how you plan to mitigate them. A common example is a low questionnaire response rate, which can be mitigated through triangulation (using multiple data sources or methods to cross-verify findings). Limitations vary from study to study and depend heavily on context; be specific rather than generic, and always pair each limitation with a mitigation strategy.

Chapter 3 Structure Template (At a Glance)

A typical research methodology chapter follows this order:

1. Introduction to the chapter
2. Research design (and justification)
3. Population of the study
4. Sample size and sample size determination
5. Sampling technique(s) and justification
6. Data collection methods and tools
7. Validity and reliability of instruments
8. Pilot study (if applicable)
9. Ethical considerations
10. Data analysis methods and tools/software
11. Limitations of the study

Following this order matters because examiners read proposals expecting this sequence. A chapter that jumps around (e.g. discussing analysis before sampling is settled) reads as disorganised even when the content is technically sound.

Common Mistakes That Get Chapter 3 Sent Back

• No justification for the chosen design. Stating the design without connecting it to the research objectives.
• Sample size with no formula or reasoning. A number with no visible working behind it.
• Sampling technique mismatch. Using random sampling language while describing a clearly purposive selection process, or vice versa.
• Missing validity and reliability section. Especially common and one of the first things examiners flag.
• Vague ethical considerations. A single sentence (“ethics will be considered”) instead of a concrete plan for consent, confidentiality, and clearance.
• Data analysis techniques that don’t match the research questions. Choosing a regression model, for instance, when the objectives call for a comparison between groups.
• Generic limitations. “Time and resources were limited” without any mitigation plan.

Frequently Asked Questions

How long should Chapter 3 be? Length varies by institution, but most Chapter 3s run 15–25 pages once all sections — design, population, sampling, data collection, validity, ethics, and analysis — are adequately covered. Depth matters more than page count.

Can I change my methodology after proposal defense? Yes, though it typically requires supervisor and sometimes committee approval, and should be reflected in an updated proposal or a formal amendment, depending on your institution’s process.

Do I need both quantitative and qualitative data? Only if your research questions genuinely require both. Mixed-methods should be a deliberate design choice tied to your objectives, not a way to appear more rigorous.

What is the difference between a research design and a research methodology? Research design refers to the overall strategy (quantitative, qualitative, or mixed) chosen to integrate the different components of the study. Research methodology is the broader term covering the full set of methods and procedures used, including design, population, sampling, data collection, and analysis.

Final Thoughts on Writing Chapter 3 of a PhD Thesis Proposal

Chapter 3 is informed directly by the research problem and research questions specified in Chapter 1. Every choice here — design, sample, tools, analysis — should trace back to those objectives in a way an examiner can follow without having to ask. Think through your study carefully from the beginning, because what’s in your introduction chapter shapes everything that follows, all the way through to your final thesis.

Free download

Related posts

How to Write Chapter 1 (Introduction) of a PhD Thesis Proposal (Complete Guide + Examples)

How to Write the Literature Review Chapter of a PhD Thesis Proposal (A Beginner’s Guide)

Ethical Use of Generative AI in PhD Research: A Practical Guide for Students