Should your thesis measure something, interpret something, or do both? Most master’s students reach the qualitative vs quantitative decision at the proposal stage with a half-formed topic and a strong preference for one approach, usually the one their department teaches most. That is the wrong order. The approach follows from the question, and a question that asks “how much” or “whether” needs different data from one that asks “how” or “why”.
This article is part of our guide How to Write a Master’s Thesis by the Mastermind PhD team and relates to our Research Proposal support service. Sources last checked: November 2026.
This guide compares the three approaches on purpose, data, sample, analysis and quality criteria, gives four decision questions you can answer in an afternoon, walks through three worked examples, and shows how to justify the choice so that an examiner reads it as reasoning rather than habit.
What is the difference between qualitative, quantitative and mixed methods research? Quantitative research collects numerical data from many cases and analyzes it statistically to measure levels, test hypotheses or estimate relationships. Qualitative research collects words, images or observations from fewer cases and interprets them to understand meanings, processes and contexts. Mixed methods research combines both strands in a single study with an explicit plan for how the two sets of results will be integrated. The labels describe the logic of inquiry and the type of data, not the rigor of the work.
Qualitative vs quantitative vs mixed methods: what does each approach do?
The three approaches differ in what they treat as evidence and in what counts as a good answer. The table below summarizes the qualitative vs quantitative contrasts that matter when you plan a thesis, with mixed methods alongside.
| Dimension | Quantitative | Qualitative | Mixed methods |
|---|---|---|---|
| Typical question words | How much, how many, whether, to what extent | How, why, what is it like | Both, in a planned sequence or in parallel |
| Logic | Deductive: theory, hypotheses, test | Inductive: data, patterns, concepts | Pragmatic: whatever combination answers the question |
| Data | Numbers from questionnaires, tests, records, experiments | Transcripts, field notes, documents, images | Both types, linked at a defined point |
| Sample | Large, ideally probability-based; often 100 to 400 for a survey | Small, purposive; often 8 to 25 interviews | Two samples, or one sample studied twice |
| Analysis | Descriptive and inferential statistics | Thematic, content, narrative or discourse analysis | Separate analyses plus integration (joint display, meta-inference) |
| Quality criteria | Validity, reliability, generalizability | Credibility, transferability, dependability, confirmability | Both sets, plus the quality of the integration |
| Typical output | Tables of means, tests and effect sizes | Themes with quotations, models of a process | Joint displays showing where the strands converge or diverge |
The qualitative quality criteria in the table are the four proposed by Lincoln and Guba (1985) as counterparts to validity and reliability; most qualitative methodology chapters still use them.
Which research questions suit which approach?
Write your main question on one line and underline its verb. On the qualitative vs quantitative choice, the verb tells you more than any textbook.
| If your question asks… | Example | Best fit |
|---|---|---|
| What is the level of X? | What is the level of job satisfaction among nurses in public hospitals? | Quantitative (descriptive survey) |
| Is there a difference or a relationship? | Does burnout differ between nurses on day and night shifts? | Quantitative (comparative or correlational) |
| Does X cause or predict Y? | Does a mentoring program reduce first-year turnover? | Quantitative (experimental or quasi-experimental) |
| How do people experience or make sense of X? | How do newly qualified nurses experience their first night shifts? | Qualitative (interviews; phenomenological or thematic) |
| How does a process unfold in context? | How does a ward team adapt to a new handover protocol? | Qualitative (case study, observation) |
| How much, and why? | How widespread is burnout, and why do some units escape it? | Mixed methods (explanatory sequential) |
| What should we measure, then how common is it? | What do nurses mean by “unsafe staffing”, and how often do they report it? | Mixed methods (exploratory sequential) |
A thesis can hold two or three sub-questions of different kinds, but each sub-question still needs the data type that can answer it. The common qualitative vs quantitative mismatch is a “how do students perceive” question answered with a five-point Likert questionnaire, which produces a level, not a perception.
📱 Discuss your proposal idea with a specialist — You are told the specialist’s field and degree before work starts, and a meeting is scheduled to review your department’s requirements.
What are the four decision questions?
Answer these in order and stop when one of them rules an approach out; together they settle the qualitative vs quantitative question for most theses.
1. What kind of answer will satisfy the question? If the answer is a number, a difference or an association, you are quantitative. If the answer is an account of meaning or process, you are qualitative. If a complete answer needs both, and you can say what the second strand adds that the first cannot, you are a candidate for mixed methods.
2. How much is already known? Quantitative designs need established constructs and, ideally, validated instruments. If the literature already offers a tested scale for your variable, a survey is efficient. If the concept is new, contested or poorly defined in your context, measuring it prematurely produces precise numbers about the wrong thing; in that qualitative vs quantitative trade-off, a qualitative study comes first.
3. What data can you actually reach in the time you have? A probability sample of 300 employees needs organizational access, a sampling frame and a response rate you can defend. Twelve interviews need twelve willing participants, recorded consent and roughly 60 to 90 hours of transcription and coding. Mixed methods needs both, which is why a one-semester dissertation rarely has room for it unless one strand is small and tightly scoped.
4. What do you and your program bring? Be honest about your statistical training, your interviewing skills and your supervisor’s expertise, and read your handbook: some programs require an empirical quantitative component, others expect a qualitative case study. Choosing against the grain in the qualitative vs quantitative decision is possible, but the proposal must justify it.
How do you choose a mixed methods design?
Creswell and Plano Clark (2018) describe three core designs, and almost every master’s-level mixed methods thesis is one of them. The notation follows Morse (1991): capitals mark the dominant strand, an arrow marks a sequence, and a plus sign marks concurrent collection.
| Design | Notation | Sequence | Use it when | Integration point |
|---|---|---|---|---|
| Convergent | QUAN + QUAL | Both strands collected in the same phase, analyzed separately, then compared | You want to corroborate or complete a picture from two angles | Merging results in a joint display |
| Explanatory sequential | QUAN → qual | Survey first, then interviews with selected respondents | The numbers raise a “why” that only participants can answer | Using survey results to choose interviewees and interview questions |
| Exploratory sequential | QUAL → quan | Interviews or focus groups first, then an instrument built from the themes | The construct is undefined or culturally specific and no scale exists | Using themes to write items, then testing the instrument |
The design you choose fixes your timeline. An explanatory sequential thesis cannot start interviews until the survey is analyzed, so the survey must close by the midpoint of your writing period. A convergent design runs both strands at once and is faster, but it demands two parallel recruitment efforts.
What do three worked examples look like?
Education, 12-month MA. Question: does a peer-tutoring scheme improve reading comprehension in year 5 pupils? The verb is “improve”, the outcome is a standardized test score, and the construct is well measured. Decision: quantitative, quasi-experimental, two intact classes (n = 28 and n = 30) with pre- and post-tests, analyzed with ANCOVA using the pre-test as the covariate. A qualitative strand was considered and dropped because the student had classroom access for one term only.
Business, 6-month MSc. Question: why do small retailers in a city center abandon a loyalty app within three months? Churn figures already exist from the provider, but the “why” is unknown and the population is small. Decision: qualitative, 14 semi-structured interviews with owners who stopped using the app, analyzed with reflexive thematic analysis (Braun & Clarke, 2006). The churn data appear in the introduction as context, not as a second strand.
Nursing, 18-month part-time MSc. Question: how prevalent is moral distress among ICU nurses in two hospitals, and how do high-scoring nurses describe its sources? Decision: explanatory sequential mixed methods. Phase 1: a validated moral distress scale sent to 240 nurses, 162 returned (a 68% response rate). Phase 2: 10 interviews with respondents in the top quartile, selected using phase-1 scores. Integration: a joint display pairing each statistical pattern (for example, higher scores on night rotation) with the themes that explain it.
In each case the qualitative vs quantitative decision was made last, after the question, the state of knowledge, access and time had been settled.
What does each approach cost in time and sample?
| Resource | Quantitative survey | Qualitative interviews | Mixed (explanatory sequential) |
|---|---|---|---|
| Participants | 150 to 400, with the sample size calculated in advance | 8 to 25, until themes stabilize | Survey sample plus 6 to 15 interviewees |
| Instrument work | Adopt or adapt a validated scale; pilot with 20 to 30 | Interview guide; pilot with 1 or 2 | Both, in sequence |
| Data collection | 3 to 6 weeks | 4 to 8 weeks including scheduling | 10 to 16 weeks |
| Analysis | 2 to 4 weeks in SPSS, R or jamovi | 6 to 10 weeks of transcription and coding | Both, plus about 2 weeks for integration |
| Main risk | Low response rate, unmet test assumptions | Thin data, description without interpretation | Running out of time before integration |
The interview figures are consistent with Guest, Bunce and Johnson (2006), who found that in a relatively homogeneous group most codes appeared within the first 12 interviews; heterogeneous samples need more.
How do you justify the qualitative vs quantitative choice in the methodology chapter?
Examiners do not want a page on positivism versus interpretivism. They want three to five sentences that link the question to the data and to the analysis, with the alternatives named and dismissed for a reason. A template:
“The study adopted a quantitative, cross-sectional survey design because the research questions ask about the level of X and its association with Y in a population of about 2,000 employees, for which a validated instrument exists (Author, year). A qualitative design was considered but would not have allowed the association in RQ2 to be estimated; a mixed methods design was not feasible within the 12-week dissertation period. The design’s limitations, including its inability to establish causation, are discussed in section 3.8.”
Replace the nouns and the paragraph justifies any qualitative vs quantitative or mixed methods choice.
Which qualitative vs quantitative mistakes do examiners flag?
- Calling a study “mixed methods” because a questionnaire had two open-ended questions. Without a planned integration, that is a quantitative study with some qualitative comments.
- Choosing qualitative methods to avoid statistics, then reporting themes as percentages of participants, which is a weak quantitative study in disguise.
- Testing hypotheses with eight interviews, or claiming generalization from a convenience sample of 60.
- Writing a philosophy section that does not change a single methodological decision.
- Selecting a sequential mixed design in a six-month program and never reaching the second strand.
- Describing the approach in the proposal as “both qualitative and quantitative” without saying which comes first, which dominates and where the two meet.
📱 Ask on WhatsApp now — 5 years of experience across all disciplines and 14 academic services. Two-stage payment (80/20); revisions are free until supervisor approval, within 30% of the chapter.
Frequently asked questions
Is mixed methods better than a single approach?
No. It answers questions that need both kinds of evidence and costs roughly the time of two studies. If your question is fully answerable with one data type, a single, well-executed approach scores higher than a thin mixed design.
Can a master’s thesis be purely qualitative?
Yes, in most programs, provided the handbook does not require statistical analysis. A qualitative thesis is judged on depth, reflexivity and a transparent analysis trail, not on sample size. Check the program rules before committing to either side of the qualitative vs quantitative choice.
How many interviews are enough for a qualitative master’s thesis?
Most supervisors accept 10 to 20 for a homogeneous group, following evidence that core themes stabilize around 12 interviews (Guest et al., 2006). Reflexive thematic analysis treats “saturation” cautiously, so justify the number by the richness of the data and the scope of the question rather than by a fixed rule.
Do I need a research paradigm section?
Many programs expect a short statement of philosophical position. Keep it to one or two paragraphs and connect it to actual decisions: why you measured rather than interpreted, why you treat participant accounts as constructions, and so on. Unconnected philosophy earns no marks.
Can I switch approach after the proposal is approved?
Sometimes, with your supervisor’s written agreement and, if ethics approval was granted, an amendment. A switch from quantitative to qualitative usually means new consent forms and a new ethics review, so settle the qualitative vs quantitative question carefully at the proposal stage.
Sources
- Creswell, J. W., & Creswell, J. D. (2018). Research design: Qualitative, quantitative, and mixed methods approaches (5th ed.). SAGE.
- Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting mixed methods research (3rd ed.). SAGE.
- Bryman, A. (2016). Social research methods (5th ed.). Oxford University Press.
- Johnson, R. B., & Onwuegbuzie, A. J. (2004). Mixed methods research: A research paradigm whose time has come. Educational Researcher, 33(7), 14–26. https://doi.org/10.3102/0013189X033007014
- Guest, G., Bunce, A., & Johnson, L. (2006). How many interviews are enough? An experiment with data saturation and variability. Field Methods, 18(1), 59–82. https://doi.org/10.1177/1525822X05279903
- Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. SAGE.
- Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa
Sources last checked: November 2026. This article is general educational guidance and does not replace your university regulations or your supervisor’s instructions. How we prepare our content.