Research design is the overall plan that connects your research question to the evidence you collect and the conclusions you can draw. In a South African dissertation it usually gets a short subsection in the methodology chapter, and it is where examiners check whether you understand your own study: a survey called an experiment, a case study described as a phenomenology, or causal claims from a cross-sectional design all signal a shaky foundation. This guide explains the main research design types, how to choose one, and how to justify yours in writing, with South African examples. If you are working on the methodology chapter now, Tesify helps you structure your dissertation chapter by chapter.
What research design means in a dissertation
Students meet several overlapping terms in methodology textbooks. A simple hierarchy helps:
| Level | Question it answers | Examples |
|---|---|---|
| Paradigm | What do you assume about reality and knowledge? | Positivism, interpretivism, pragmatism, critical |
| Approach | What kind of data will you use? | Quantitative, qualitative, mixed methods |
| Research design | What is the overall strategy for answering the question? | Survey, experiment, case study, phenomenology, grounded theory |
| Methods | How exactly will you collect and analyse data? | Questionnaire, interviews, regression, thematic analysis |
Your faculty may use slightly different labels. What matters is that the levels fit together and that your chapter states each one explicitly.
Quantitative research designs
- Descriptive survey (cross-sectional): measures variables at one point in time to describe levels and patterns. Example: the prevalence of food insecurity among students at one university (illustrative).
- Correlational: tests whether variables are related without manipulating anything. Example: the relationship between perceived organisational support and work engagement among municipal employees.
- Comparative (causal-comparative): compares existing groups on an outcome. Example: comparing mathematics anxiety in learners from fee-paying and no-fee schools.
- Quasi-experimental: compares groups before and after an intervention without random assignment. Example: evaluating a peer tutoring programme in two classes.
- Experimental: randomly assigns participants to conditions to test cause and effect. Less common in coursework master’s research because of time, ethics and logistics.
- Longitudinal: measures the same people more than once over time. Usually too long for a master’s, but possible with secondary panel data.
Our guide to which statistical test to use shows how each quantitative design leads to particular analyses.
Qualitative research designs
- Generic qualitative (descriptive): explores participants’ views on a practical issue without a specialised tradition. Common in health sciences and education research reports.
- Case study: examines one or a few bounded cases in depth, often using several data sources. Example: how one district implemented an inclusive education policy.
- Phenomenology: studies the lived experience of a phenomenon. Example: the experience of first-generation students in postgraduate study.
- Grounded theory: builds a theory of a process from data. Demanding for a short project.
- Narrative inquiry: studies people’s stories over time.
- Ethnography: studies a culture or setting through extended immersion.

Mixed methods designs
Mixed methods combine quantitative and qualitative strands in one study. The common designs are explanatory sequential (survey, then interviews to explain results), exploratory sequential (interviews first, then a survey built from them) and convergent (both strands at once, compared). Choose mixed methods only when the question genuinely needs both kinds of evidence and your timeline allows two full analyses plus an integration section.
How to choose your research design
- Read your research question’s verb. “Determine the level of” suggests descriptive; “examine the relationship between” suggests correlational; “compare” suggests comparative; “evaluate the effect of an intervention” suggests quasi-experimental; “explore experiences of” suggests qualitative.
- Check the claim you want to make. If you want to say X causes Y, you need an experimental or strong quasi-experimental design. If you can only say X is associated with Y, a correlational design is honest.
- Check your access and time. A longitudinal design is impossible in a one-year coursework master’s unless you use existing data.
- Check consistency. Your paradigm, approach, design and methods must point in the same direction.
- Check your supervisor’s expectations. Some departments prefer particular designs for research reports at honours or coursework master’s level.
If your aims and objectives are not yet stable, fix those first; our research aims and objectives examples show how wording leads to design.
How to justify your research design
Examiners do not want a textbook definition followed by “this design was used”. They want an argument. A strong justification has four parts:
- Name the design precisely (“a quantitative, cross-sectional, correlational survey design”).
- Define it briefly with a methodology source your faculty accepts.
- Link it to your question: why this design answers what you asked.
- Acknowledge its limitation and why you accepted it, ideally comparing it with an alternative you rejected.
An illustrative paragraph: “A quantitative, cross-sectional, correlational design was adopted because the study aimed to examine the relationship between work engagement and turnover intention among professional nurses at a single point in time. An experimental design was not feasible, as engagement cannot be manipulated ethically, and a longitudinal design was beyond the timeframe of the degree. The design does not permit causal conclusions; findings are therefore reported as associations.”

Research design and sampling go together
Your design shapes your sampling. Correlational and descriptive surveys that aim to generalise need probability sampling and a justified sample size; qualitative designs use purposive sampling and saturation. Mixing them up, for example claiming generalisable prevalence from a convenience sample, is one of the most common examiner comments. Our population and sampling examples show how to describe both properly.
Research design examples by field (illustrative)
| Field | Illustrative question | Design |
|---|---|---|
| Nursing | What factors are associated with medication errors in a provincial hospital? | Quantitative cross-sectional correlational, using records and a survey |
| Education | How do teachers in one circuit implement a reading intervention? | Qualitative multiple case study |
| Accounting | Has a reporting standard changed disclosure quality among listed companies? | Quantitative longitudinal analysis of annual reports |
| Psychology | What is the lived experience of postgraduate students with impostor feelings? | Qualitative interpretative phenomenology |
| Public health | Did a community health worker programme change clinic attendance? | Quasi-experimental pre-post design with comparison clinics |
Threats to validity: what each design must defend
Every design has typical weaknesses, and examiners expect you to name the ones that apply to yours and say what you did about them.
| Design | Main threat | What you can do |
|---|---|---|
| Cross-sectional survey | Cannot establish time order or cause; non-response bias | Use cautious language; compare early and late respondents; report response rate |
| Correlational | Confounding variables explain the relationship | Measure and control for likely confounders in regression |
| Comparative | Groups differ in ways other than the one studied | Describe group characteristics; control statistically where possible |
| Quasi-experimental | Groups not equivalent at baseline; other events during the study | Pre-test both groups; record contextual events; use a comparison group |
| Case study | Limited transferability; researcher bias | Rich description; multiple data sources; reflexive notes |
| Phenomenology | Researcher’s assumptions shaping interpretation | Bracketing or reflexivity; member checking |
Worked example: revising a design after supervisor feedback (illustrative)
A master’s student in business management proposed to “determine the effect of remote work on employee productivity” at a Cape Town insurer using an online questionnaire. The supervisor flagged two problems: “effect” implied a causal claim the cross-sectional survey could not support, and productivity was measured only by self-report. The student changed three things:
- The aim became “to examine the relationship between remote work arrangements and self-reported productivity”, matching a correlational design.
- The design was named precisely as a quantitative, cross-sectional, correlational survey, with the causal limitation stated.
- Control variables (tenure, role, caring responsibilities) were added so the analysis could account for obvious confounders.
The study became smaller in its claims but far easier to defend, which is usually the right trade in a coursework degree.
Research design in the proposal versus the dissertation
In the research proposal, the design is written in the future tense and focuses on the plan and its justification. In the final dissertation, it is written in the past tense and must report what actually happened, including any changes from the proposal and why they were made. If fieldwork forced a change (for example, a planned comparison group became unavailable), say so openly and explain how it affects your conclusions. Examiners respect honest reporting of changes; they penalise designs that are described as planned but clearly were not carried out that way. Our guide on how to write a research proposal for a South African university covers the proposal version.
Paradigm and design: making them fit
Many South African supervisors ask for the paradigm before the design. Keep the link simple. A positivist or post-positivist paradigm fits quantitative designs that measure and test; an interpretivist or constructivist paradigm fits qualitative designs that explore meaning; pragmatism is often used to justify mixed methods because it prioritises what works for the question. State your paradigm in a short paragraph and then show, in one sentence, why your design follows from it.
Where research design sits in your chapter
Most South African methodology chapters follow an order close to this: introduction, research paradigm, research approach, research design, population and sampling, data collection, data analysis, quality criteria (validity and reliability, or trustworthiness), ethical considerations, limitations and summary. Our guide on writing the methodology chapter of a nursing master’s dissertation shows the full structure in one field.
A research design checklist before you submit the chapter
- The design is named precisely, with approach, time dimension and type (for example, quantitative cross-sectional correlational).
- A short definition is supported by a methodology source your faculty accepts.
- The justification links the design to your aim and research questions in your own words.
- At least one alternative design is mentioned and the reason for rejecting it is given.
- The main limitation of the design is acknowledged, and your conclusions respect it.
- Paradigm, approach, design, sampling and analysis are consistent with each other.
- Any change from the approved proposal is reported and explained.
Common research design mistakes
- Using design and method interchangeably (“the design was a questionnaire”).
- Claiming an experimental design without random assignment.
- Causal language in a correlational study.
- Calling a study phenomenological while using a generic interview approach and thematic analysis without a phenomenological stance.
- A mixed methods label with no integration of the two strands.
- Paradigm and design that contradict each other.
FAQ
What is research design in simple terms?
It is the overall strategy for answering your research question: what kind of study you will run, with whom, and what conclusions it allows.
What is the difference between research design and methodology?
Methodology is the whole account of how and why you did the research. Research design is one part of it: the overall strategy, such as a survey or case study.
Which research design is best for a master’s dissertation?
The one that fits your question, access and timeline. Cross-sectional surveys and generic qualitative or case study designs are the most common in coursework master’s research.
Can a correlational design show cause and effect?
No. It shows association. Use cautious language such as “was associated with” or “predicted”.
How long should the research design section be?
Usually one to two pages: name, define, justify, acknowledge limitations. Follow your supervisor’s guidance.
Does Tesify choose my research design?
No. Tesify helps you structure your dissertation chapter by chapter; the research decisions and the text stay yours, 100% written by you.
Structure your methodology with Tesify
A methodology chapter reads well when every level, from paradigm to methods, follows from your question. 50,000+ students have written 230,000+ chapters with Tesify, which helps you structure your dissertation chapter by chapter, and every word stays 100% written by you. Structure your dissertation with Tesify.
