Qualitative vs quantitative research is the most fundamental choice you make in your methodology. Quantitative research measures variables as numbers and tests hypotheses statistically to find patterns that generalise to a population; qualitative research explores meaning, experience and context through words, building a rich interpretation of how and why something happens. Put simply: quantitative research asks how many and how strongly; qualitative research asks how and why.
Neither is superior. The right approach depends on your research question, your assumptions about knowledge, and what you need your findings to do. This guide defines both, compares them across every dimension a marker cares about, walks through two parallel worked studies on the same topic, explains mixed methods, and gives you a practical framework for choosing.
What is quantitative research?
Quantitative research is the systematic investigation of phenomena by collecting numerical data and analysing it with statistics. It is built on the philosophy of positivism: the assumption that there is an objective reality that can be measured independently of the researcher. The researcher typically begins with a theory, derives a testable hypothesis, operationalises concepts into measurable variables, collects data from a sample, and uses statistical tests to decide whether the data support the hypothesis. Because the logic moves from a general theory down to a specific prediction that is then tested, quantitative work is usually deductive (see our guide to inductive and deductive reasoning).
Typical quantitative designs include experiments, quasi-experiments, surveys with closed questions, and correlational studies. Outputs are numbers: means, percentages, correlation coefficients, p-values, effect sizes and confidence intervals. The prized qualities are reliability (consistency of measurement), internal validity (the design really isolates the relationship claimed), and generalisability (results extend from the sample to the wider population).
What is qualitative research?
Qualitative research investigates how people understand and experience the world, generating non-numerical data such as interview transcripts, field notes, documents and images. It is grounded in interpretivism (and related positions such as constructivism): the view that social reality is constructed through meaning, so the researcher’s job is to interpret participants’ perspectives in context rather than to measure them. The logic is usually inductive — patterns, concepts and theory are built up from the data rather than imposed in advance.
Common qualitative designs include phenomenology, grounded theory, ethnography, case study and narrative inquiry. Data are collected through in-depth interviews, focus groups, observation and document analysis, and analysed through approaches such as thematic analysis or content analysis. The prized qualities here are not reliability and generalisability but trustworthiness — credibility, transferability, dependability and confirmability (Lincoln & Guba, 1985) — achieved through techniques like member checking, thick description and an audit trail.
“Qualitative research is a means for exploring and understanding the meaning individuals or groups ascribe to a social or human problem… Quantitative research is a means for testing objective theories by examining the relationship among variables.” (Source: Creswell & Creswell, 2018)
The master comparison: qualitative vs quantitative research
The clearest way to grasp the distinction is to lay the two approaches side by side across every dimension that shapes a dissertation methodology. Read this table as a checklist: each row is a decision you will have to justify to your marker.
| Dimension | Quantitative | Qualitative |
|---|---|---|
| Aim | Test hypotheses; measure variables; quantify relationships and generalise | Explore meaning, experience and process; understand the “how” and “why” |
| Data type | Numbers (scores, counts, ratings, measurements) | Words and images (transcripts, field notes, documents) |
| Reasoning | Mostly deductive (theory → hypothesis → test) | Mostly inductive (data → patterns → theory) |
| Philosophy | Positivism / post-positivism; objective reality | Interpretivism / constructivism; socially constructed reality |
| Sampling | Probability sampling (random, stratified) for representativeness | Purposive, theoretical or snowball sampling for information richness |
| Sample size | Large (often 100s) for statistical power | Small (often 6–30) to data saturation |
| Data collection | Closed surveys, experiments, structured instruments, secondary datasets | In-depth interviews, focus groups, observation, documents |
| Analysis | Statistical: descriptive & inferential (t-tests, ANOVA, regression) | Interpretive: thematic, content, narrative, grounded-theory coding |
| Researcher role | Detached, objective; minimise influence on data | Reflexive instrument; positionality acknowledged |
| Quality criteria | Reliability, validity, generalisability | Credibility, transferability, dependability, confirmability |
| Typical output | Tables, statistics, charts; supported/rejected hypotheses | Themes, categories, models; rich narrative with quotations |
When to use quantitative research
Choose a quantitative approach when your question is about measurement, prediction or testing a relationship that you can express numerically. It fits naturally with the questions a survey or experiment can answer, and it lets you make claims that generalise beyond your sample. Use it when:
- You want to test a clear, pre-specified hypothesis (e.g. “does X cause Y?” or “is X associated with Y?”).
- Your concepts can be operationalised into reliable, measurable variables (see types of research for design families).
- You need results that generalise to a population and can be replicated.
- You have access to a large enough sample for adequate statistical power.
- Comparison, ranking or correlation between groups or variables is the point.
When to use qualitative research
Choose a qualitative approach when you need depth, context and the participant’s own framing — especially where little is known and a rigid hypothesis would be premature. Use it when:
- You want to understand lived experience, meaning or process (“what is it like to…”, “how do people make sense of…”).
- The topic is under-researched and you are generating, not testing, theory.
- Context and nuance would be flattened by reducing the phenomenon to numbers.
- You are studying a small, hard-to-reach or unique group where rich detail matters more than breadth.
- You expect the important variables to emerge from the data rather than being known in advance.
Strengths and limitations of each approach
Every design buys some advantages at the cost of others. Knowing the trade-offs lets you defend your choice and pre-empt your examiner’s objections.
Quantitative: strengths
- Results generalise to the population when sampling is sound.
- Objective, replicable and comparatively quick to analyse at scale.
- Statistical tests quantify the strength and significance of relationships (see hypothesis testing).
- Large samples reduce the influence of outliers and chance.
Quantitative: limitations
- Strips out context and meaning — the “why” behind the numbers stays hidden.
- Only as good as the instrument: a poorly worded survey yields precise but meaningless data.
- Rigid design cannot adapt to unexpected findings mid-study.
- Risk of false precision and over-claiming causation from correlation.
Qualitative: strengths
- Rich, detailed, contextualised understanding of how and why.
- Flexible — the researcher can follow up and probe in real time.
- Excellent for theory generation and exploring sensitive or complex experiences.
- Captures the participant’s own voice and meaning.
Qualitative: limitations
- Findings rarely generalise statistically to a wider population.
- Labour-intensive: transcription and coding are slow.
- More open to researcher bias, so reflexivity and an audit trail are essential.
- Smaller samples and interpretive judgement make replication harder.
Two parallel worked examples on the same topic
The fastest way to feel the difference is to take one topic — the effect of remote working on employees — and design it both ways. Notice how the question, sample, data and analysis diverge end to end even though the subject is identical.
Question: Is the number of remote days per week associated with employee job satisfaction?
Hypothesis (H1): There is a positive correlation between remote days/week and a job-satisfaction score.
Sample: 120 office employees, recruited by stratified random sampling across departments.
Data: A validated 7-item satisfaction scale (each item 1–7) plus a single item for remote days/week (0–5). Each respondent’s satisfaction score = sum of 7 items (range 7–49).
Analysis — Pearson’s r, worked on a 5-person illustration:
Remote days (x): 1, 2, 3, 4, 5 → mean x̄ = (1+2+3+4+5)/5 = 3
Satisfaction (y): 28, 30, 35, 38, 44 → mean ȳ = (28+30+35+38+44)/5 = 175/5 = 35
Deviations (x−x̄): −2, −1, 0, 1, 2 | (y−ȳ): −7, −5, 0, 3, 9
Products (x−x̄)(y−ȳ): 14, 5, 0, 3, 18 → Σ = 40
Σ(x−x̄)² = 4+1+0+1+4 = 10; Σ(y−ȳ)² = 49+25+0+9+81 = 164
r = 40 / √(10 × 164) = 40 / √1640 = 40 / 40.50 = 0.988
Interpretation: r ≈ 0.99 is a very strong positive correlation, so on this illustrative data we would reject the null hypothesis (subject to the p-value and full n=120 sample). Caveat: correlation is not causation — a confound such as seniority could drive both. The deliverable is a number with a confidence interval that generalises to the population.
Question: How do employees experience the shift to remote working, and what shapes their sense of satisfaction?
Sample: 12 employees chosen by purposive sampling for maximum variation (role, tenure, home setup), recruited to data saturation.
Data collection: 45–60 minute semi-structured interviews, audio-recorded and transcribed verbatim.
Analysis — reflexive thematic analysis (Braun & Clarke, 2006), six phases:
1) Familiarisation — read all 12 transcripts twice. 2) Coding — tag segments, e.g. “blurred work–home boundary”, “reclaimed commute time”, “loss of spontaneous chat”. 3) Searching for themes — cluster codes. 4) Reviewing. 5) Defining. 6) Writing up.
Result — three themes: (i) Autonomy as a double-edged gift; (ii) The vanishing boundary between work and home; (iii) Belonging at a distance. Each theme is evidenced with anonymised participant quotations.
Interpretation: satisfaction is not a single score but a negotiated balance between freedom and isolation — an insight a correlation could never surface. The deliverable is a rich, contextual model of why remote working affects people, transferable rather than statistically generalisable.
Same topic; two entirely different studies. The quantitative version tells you that more remote days track with higher satisfaction and how strongly; the qualitative version tells you what that satisfaction actually consists of and why it can flip into isolation. This is exactly why researchers increasingly combine them.
Mixed methods research
Mixed methods research integrates quantitative and qualitative data within a single study to draw on the strengths of both. Creswell & Plano Clark argue that combining the two yields a more complete understanding than either alone, provided the integration is deliberate rather than two studies stapled together. The three core designs you should know are:
| Design | Sequence | Purpose |
|---|---|---|
| Convergent parallel | QUAN + QUAL collected together, analysed separately, then merged | Corroborate and compare two views of the same problem at once |
| Explanatory sequential | QUAN → then QUAL | Use qualitative follow-up to explain surprising or significant quantitative results |
| Exploratory sequential | QUAL → then QUAN | Use early qualitative work to build an instrument or hypotheses to then test at scale |
In our remote-working example, an explanatory sequential design would first run the survey on 120 staff, then interview a purposive subset of high- and low-satisfaction respondents to explain why the correlation appears. An exploratory sequential design would reverse it: interview first to discover what “satisfaction” means to remote staff, then build and test a tailored survey. A convergent design would run both at once and merge them.
How to choose: a step-by-step framework
Work through these steps in order; do not start from the method you find comfortable.
- Start with the research question. “How many / how much / is X related to Y” points quantitative; “how / why / what is it like” points qualitative.
- Check the state of knowledge. A well-mapped field with established variables supports testing (quantitative); a new or poorly understood phenomenon supports exploring (qualitative).
- Examine your philosophical position. Do you believe the thing can be objectively measured (positivism) or that it is meaningfully constructed (interpretivism)?
- Match reasoning to aim. Testing a theory is deductive; building one is inductive.
- Audit feasibility. Can you reach a large random sample, or only a few rich informants? Consider time, access and your own analytical skills.
- Decide single or mixed. If one method leaves an important part of the question unanswered, choose a mixed design and state the integration point explicitly.
For deeper detail on the techniques behind each side, see our guides to methods of data collection, the qualitative research questionnaire and the quantitative research questionnaire.
Common mistakes to avoid
- Choosing the method first and bending the question to fit it — the question must lead.
- Calling a study “mixed methods” when you simply added one open question to a survey; true mixing requires genuine integration.
- Applying quantitative quality criteria (generalisability, large n) to a qualitative study, or vice versa.
- Claiming causation from a correlation, or claiming generalisability from 12 interviews.
- Ignoring philosophy — markers expect you to justify your positivist or interpretivist stance.
- Treating a tiny qualitative sample as “too small” — saturation, not size, is the standard.
Not sure which approach fits your dissertation?
Our academic experts help you choose and justify the right methodology, design your study and analyse your data — quantitative, qualitative or mixed.
Related methodology guides
- Mixed Methods Research
- Research Philosophy
