The chain of justification
A good methodology chapter reads as a chain: the question determines the approach, the approach determines the design, and the design determines the sample, data collection and analysis. Each link should be justified in a sentence or two, with a citation to a methods text. If one link is missing, the chapter reads as a list.
| Layer | Choices | Question to answer |
|---|---|---|
| Philosophy | Positivism, interpretivism, pragmatism, critical realism | What counts as knowledge in this study? |
| Approach | Deductive, inductive, abductive | Am I testing theory or building it? |
| Design | Survey, experiment, case study, ethnography, interviews, mixed | What strategy best answers the question? |
| Time horizon | Cross-sectional, longitudinal | One snapshot or change over time? |
| Data collection | Questionnaire, interviews, observation, documents, secondary data | What data will I gather and how? |
| Analysis | Statistics, thematic analysis, content analysis | How will I turn data into answers? |
Philosophy without the mystery
Research philosophy sounds abstract, but it only asks what kind of evidence you trust. State your position briefly and make sure it is consistent with your design.
| Position | Belief | Typical methods |
|---|---|---|
| Positivism | Reality is objective and measurable; knowledge comes from testing hypotheses | Surveys, experiments, statistics |
| Interpretivism | Reality is socially constructed; understand meaning from the participants' view | Interviews, observation, thematic analysis |
| Pragmatism | The question decides; use whatever methods best answer it | Mixed methods |
| Critical realism | A real world exists but is known imperfectly through social lenses | Case studies, mixed methods |
Deductive research starts from theory, derives hypotheses and tests them with data. Inductive research starts from data and builds themes or theory. Most dissertations are one or the other, and some move between them.
Choosing the design
| Design | Best for | Strength | Limitation |
|---|---|---|---|
| Survey | Measuring attitudes or behaviors across many people | Generalizable, efficient | Shallow; depends on response rate and good items |
| Experiment | Testing cause and effect | Strong on causation | Artificial setting; ethics; sample often students |
| Case study | How and why in context | Rich detail, multiple sources | Limited generalization |
| Semi-structured interviews | Experience, meaning, decision processes | Depth, flexibility | Time-consuming; interviewer effects |
| Secondary data analysis | Using existing datasets or reports | Saves time; large samples | Data may not fit your question |
| Mixed methods | Questions needing both breadth and depth | Triangulation | Heavy workload; integration is hard |
Do not choose mixed methods only to look rigorous. Use them when one method cannot fully answer the question, and say how the strands will be combined, for example interviews to explain surprising survey results.
Sampling and sample size
The sampling approach should match the design. Probability sampling (simple random, stratified, cluster) allows statistical generalization. Non-probability sampling (convenience, purposive, snowball) suits qualitative work and cases where a sampling frame does not exist, but findings cannot be generalized statistically.
| Approach | How it works | Use when |
|---|---|---|
| Simple random | Every member has an equal chance | You have a full list of the population |
| Stratified | Random samples within subgroups | Subgroups matter and must be represented |
| Purposive | Choose people with relevant experience | Qualitative studies needing information-rich cases |
| Snowball | Participants refer others | Hard-to-reach populations |
| Convenience | Whoever is available | Pilot studies; weak for generalizing |
State the target population, the sampling frame, the method, the size and your reasoning. Quantitative sample sizes are calculated (see our sampling guide); qualitative sample sizes are justified by information power and saturation, commonly 10 to 30 interviews for a focused study.
Validity, reliability and trustworthiness
Every study must say how it protects the quality of its findings, and the language depends on the tradition.
| Quantitative term | Meaning | How to address |
|---|---|---|
| Reliability | Consistent results; items measure the same thing | Cronbach's alpha (commonly 0.70 or higher is acceptable); test-retest |
| Validity | Measures what it claims to | Use validated scales; pilot; expert review |
| Generalizability | Results apply beyond the sample | Probability sampling; adequate size |
| Qualitative term | Meaning | How to address |
|---|---|---|
| Credibility | Findings reflect participants' realities | Member checking; triangulation; prolonged engagement |
| Transferability | Readers can judge fit to other settings | Thick description of context |
| Dependability | The process is documented and consistent | Audit trail; decision log |
| Confirmability | Findings come from data, not researcher bias | Reflexivity; second coder |
Matching question to method
| If your question asks | Consider | Typical analysis |
|---|---|---|
| How many, how much, how often | Survey or secondary data | Descriptive statistics |
| Is there a relationship between X and Y | Survey; correlational design | Correlation, regression |
| Does X cause Y | Experiment or quasi-experiment | Group comparison, controlled analysis |
| How do people experience or understand X | Interviews or focus groups | Thematic analysis |
| How and why did X happen in this organization | Case study | Within-case and cross-case analysis |
| Why does a survey result occur | Mixed methods: survey then interviews | Statistics plus themes, integrated |
Check the match in both directions: could this method produce an answer to the question, and could this question be answered better another way? Examiners probe exactly here.
Working on this assignment now? Get a price for help with your paper.
Get an instant quoteMixed methods designs
| Design | Sequence | Use when |
|---|---|---|
| Explanatory sequential | Survey first, then interviews to explain results | You expect surprising or unclear statistical results |
| Exploratory sequential | Interviews first, then a survey built from what you learned | You need to develop measures or test themes at scale |
| Convergent | Both at the same time, then compared | You want to cross-check findings from different sources |
Say how the strands will be integrated, for example by using interview findings to explain regression results, and be realistic about workload: two methods mean two sets of ethics, sampling, analysis and write-up.
A model justification paragraph
Justifying a survey design (hypothetical)
The study asks to what extent supervisor support and onboarding quality predict intention to stay. Because the aim is to estimate the strength of relationships across a population of relationship managers, a cross-sectional survey with validated scales is appropriate (cite the methods text your course uses). Interviews would give depth but could not estimate the size of the effects, and an experiment is impractical because managers cannot ethically be assigned to poor onboarding. The survey's main limitation is that it cannot establish causation, which is acknowledged in Chapter 6.
Notice the structure: aim, design, reason, alternatives considered and the limitation.
Ethics and data management in the method chapter
Methodology chapters should include a short ethics section: approval number and body, how consent was obtained, how identities are protected and how data are stored and deleted. Mention any pressure points, such as studying your employer or interviewing subordinates, and how you reduced them.
| Item | Typical statement |
|---|---|
| Approval | The study was approved by the university research ethics committee (reference given) |
| Consent | Participants received an information sheet and gave written or recorded consent |
| Anonymity | Names were replaced with codes; organization names are disguised |
| Data handling | Files stored on encrypted university storage; deleted after the retention period |
Limitations by design
| Design | Typical limitation | Mitigation to describe |
|---|---|---|
| Cross-sectional survey | Cannot show cause; common method bias | Careful wording; separate scales; statistical checks; temper claims |
| Convenience sample | Cannot generalize | Describe the sample; compare with population; state boundaries |
| Interviews | Researcher influence; small sample | Reflexive journal; second coder; purposive variety |
| Case study | Limited generalization | Rich description; link to theory (analytic generalization) |
| Secondary data | Variables not designed for your question | Justify proxies; test robustness |
Put the limitation and its mitigation side by side: it shows that you saw the weakness and acted on it.
Writing the chapter
- Justify each choice Say why, not only what, with a citation.
- Match to the question Show the logic from question to method.
- Be precise about the procedure A reader should be able to replicate your study.
- Cover ethics and limits And what you did to reduce them.
- Do not narrate the history of your changes Describe the final design, with major adjustments noted briefly.
For data analysis chapters, see our guides to thematic analysis and the data analysis chapter. If you want help with a methodology chapter, you can order graduate business research paper help.