Quantitative vs Qualitative Research: Which Should Your Thesis Use? (2026)
| Quantitative | Qualitative | |
|---|---|---|
| Answers | How many, how much, how often, is there a difference or relationship | How, why, what does it mean, what is the experience |
| Data | Numbers | Words, observations, documents |
| Typical instrument | Survey questionnaire, standardised scale, existing records | Interview guide, focus group guide, observation protocol |
| Respondents | Many; sample size calculated | Few; recruited until themes repeat |
| Sampling | Usually random or stratified | Usually purposive |
| Analysis | Statistical treatment — frequencies, means, t-test, ANOVA, correlation | Thematic analysis, coding, categories |
| Heaviest workload | Distribution and encoding | Transcription and coding |
| Main risk | Too few returns to run your planned test | Underestimating transcription and analysis time |
| Result | Generalisable within your population | Deep, contextual, not generalisable |
The choice is not yours to make
This is the part students resist, so it is worth saying first: your research question determines your design. You do not pick the design you are comfortable with and then write questions to fit it.
Look at the questions in your statement of the problem and read the verbs.
- “What is the level of…”, “Is there a significant difference between…”, “Is there a relationship between…” — these are quantitative questions. They require numbers and a statistical test.
- “How do students experience…”, “Why do teachers…”, “What meanings do…” — these are qualitative questions. They require words and thematic analysis.
If your questions and your intended design disagree, one of them is wrong. Fix it in Chapter 1, before your proposal is locked, because discovering the mismatch after data collection means starting again.
Ranked for a Philippine thesis on a semester timeline
1. Descriptive quantitative — the safest default
Who it suits. Undergraduate groups with one semester, a defined population they can actually reach, and questions about levels, frequencies and differences between groups.
Why it ranks first. The timeline is predictable. A questionnaire can be distributed to many respondents at once, encoding is mechanical, and the analysis is bounded — you know in advance which tests you will run. Panels are also thoroughly familiar with the format, so your Chapter 3 will be judged against a clear standard rather than an unusual one.
Where it falls short. It tells you what, not why. If your interesting finding is that one group scores lower, a purely quantitative study cannot explain the reason, and a panelist will ask.
2. Qualitative — better answers, harder schedule
Who it suits. Studies of experience, process or meaning; topics where little local research exists; and groups whose population is too small for meaningful statistics.
Why it places second despite producing richer findings. The workload is back-loaded and consistently underestimated. An hour of interview commonly takes several hours to transcribe, and coding takes longer still. Ten interviews is not a small job; it is most of a semester.
Where it falls short. Findings do not generalise, and you must say so rather than implying they do. Purposive sampling also has to be justified carefully — “these were the people who agreed” is not a sampling strategy.
3. Mixed methods — powerful, and usually too much
Who it suits. Students with a genuinely two-part question, a realistic timeline, and an adviser who has supervised mixed methods before.
Where it falls short. You are doing two studies. You need both instruments validated, both samples recruited, both analyses run, and a defensible account of how the two integrate. Undergraduate groups regularly propose mixed methods in September and quietly abandon one half by January — which then leaves a Chapter 3 describing work that was never done.
Choose it only if you can name, right now, exactly how the qualitative phase informs or explains the quantitative one.
What your adviser will ask when you propose a design
Design proposals get sent back for a small number of predictable reasons, and you can pre-empt every one of them.
“Who exactly are your respondents, and can you reach them?” This is the first question and the one that kills the most proposals. Name the population, say how many there are, and say how you will obtain access. If your study involves minors or a school other than your own, access means written permission from a gatekeeper, and obtaining it takes weeks rather than days.
“Where did your instrument come from?” Either you adopted a published one — in which case cite it and secure permission where required — or you wrote your own, in which case expect to have it validated by experts and to report a reliability figure. There is no third option, and “I made a questionnaire” is not an answer.
“Which test answers which question?” For quantitative designs, an adviser expects a direct mapping: research question two is answered by the weighted mean, research question three by an independent samples t-test, and so on. If you cannot produce that mapping, the design is not finished.
“What will you do if returns are low?” Having an answer — extending the collection window, broadening to a second section, adjusting the analysis — shows you have planned for the most likely failure rather than assuming it away.
The recommendation
Write the research questions first, then let them choose. If they are genuinely open and you are on a single-semester undergraduate timeline, choose descriptive quantitative.
That is not because quantitative research is better. It is because its schedule is predictable, and the most common cause of an unfinished undergraduate thesis is not weak analysis — it is running out of time during data collection.
If your question is genuinely a “why” or “how” question, do the qualitative study and plan the transcription time honestly from week one.
What changes in Chapter 3
Your design decision reshapes the methodology chapter substantially, so know what you are signing up for.
A quantitative Chapter 3 must state the population, the sampling technique, how the sample size was determined, the instrument and where it came from, its validity and reliability, the data gathering procedure, and the statistical treatment — naming which test answers which research question.
A qualitative Chapter 3 must state the design, the participants and why they were selected, the interview or observation guide and how it was developed, how data were recorded and transcribed, the analysis approach, and how trustworthiness was addressed — credibility, transferability, dependability and confirmability, or your manual’s equivalent.
Note what is common to both: you must justify your instrument. Borrowing a published scale means citing it and, where required, obtaining permission. Writing your own means having it validated by experts and, for quantitative work, reporting a reliability figure. Neither route is optional, and both take longer than students plan for.
Three mistakes that cost whole semesters
- Choosing qualitative to avoid statistics. Thematic coding is not easier than running a t-test. It is slower and less structured, and it is harder to defend when done carelessly.
- Choosing quantitative without checking you can reach the respondents. A calculated sample of 380 means nothing if you have access to 90 people. Confirm access before you commit to the design.
- Deciding the statistical test after collecting data. The test must be named in Chapter 3 and must match the question. Collecting first and then hunting for a test that produces significance is the wrong way round, and panels ask directly which test you planned.
Frequently asked questions
Which is better for a thesis, quantitative or qualitative?
Neither. Your research question determines which is appropriate, and choosing against your question produces data that cannot answer it.
Which is easier?
Quantitative is usually more predictable to schedule. Qualitative is not easier — transcription and coding are slow and are routinely underestimated.
How many respondents do I need for a quantitative thesis?
It depends on your population and the sampling approach your manual accepts. State how the size was determined rather than asserting a number.
How many participants for a qualitative study?
Recruitment usually continues until themes repeat and new interviews stop adding anything. Your manual may set expectations, so check.
Can I do mixed methods for an undergraduate thesis?
You can, but you are running two studies on one timeline. Only choose it if you can state exactly how the two phases connect.
Can I change design after my proposal is approved?
Changing after approval usually means re-approval and possibly a repeat proposal defense. Decide before the proposal is locked.
Do qualitative studies need statistical treatment?
No, but they need a stated analysis approach and a trustworthiness section, which serve the same function of showing your analysis was systematic.
Do I need to validate my own questionnaire?
Yes if you wrote it. Expert validation and a reliability measure are standard requirements — check what your manual specifies.
Which design suits a capstone project?
Capstone evaluation is commonly quantitative, using a structured instrument with users. Our comparison of thesis and capstone requirements covers what changes.
Where do I record the design decision?
In Chapter 3, with justification — and it must be consistent with the questions in Chapter 1 and the gap in Chapter 2.
Decide once, then build
The design choice takes an afternoon if you read your own research questions honestly. What takes months is discovering in February that the data you collected cannot answer them.
Tesify keeps your questions, method and analysis aligned as you draft, so a change in one chapter surfaces where it affects the others — 100% written by you, alongside 9,000+ students and 15,000+ chapters.