Quantitative, qualitative and mixed methods research: how they differ and how to choose

This guide compares the three on data, sampling, analysis, strengths and limits, explains the three core mixed methods designs, and ends with a way to justify your choice. It is part of our types of research design series.

What is quantitative research?

Quantitative research measures variables with structured instruments and analyzes the numbers with statistics to describe, compare or test relationships. It usually starts from a hypothesis drawn from theory or earlier studies (deductive reasoning), and it aims to generalize from a sample to a population.

• Data: counts, scores, measurements, ratings, records. Each variable has an operational definition that says exactly how it is measured.
• Typical designs: experiments, quasi-experiments, surveys, correlational and cohort studies.
• Output: estimates with confidence intervals, effect sizes, test results.

Example (illustrative), business: "Is the number of days worked from home per week associated with job satisfaction scores among 500 software developers, after adjusting for age and seniority?"

What is qualitative research?

Qualitative research collects open-ended, non-numerical data and interprets it to understand experiences, meanings, processes and contexts. It often works inductively, building ideas from the data rather than testing a prediction, and it aims for depth and transferability, not statistical generalization.

• Data: interview and focus group transcripts, field notes, documents, images, open survey answers.
• Typical designs: case study, ethnography, grounded theory, phenomenology, narrative research, qualitative description.
• Output: themes, categories, a theory or model, a rich description, supported by quotations.

Example (illustrative), health: "How do community health workers in rural districts describe the barriers they face when following up patients who miss tuberculosis treatment?"

What is mixed methods research?

Mixed methods research collects and analyzes both quantitative and qualitative data within one study or program and integrates them, so the combination answers more than either strand would alone. Johnson, Onwuegbuzie and Turner reviewed definitions from leading methodologists and found that integration of the two kinds of data, not just their presence, is what defines it [1].

• Data: both, with a plan for where they meet.
• Integration: by merging results, by using one phase to build or select for the next, or by embedding one inside the other [2].
• Output: findings from each strand plus a combined interpretation, often shown in a joint display (a table with numbers and themes side by side).

Example (illustrative), education: a survey of 600 students measures how often they use the library's study skills service and their grades; interviews with 20 frequent and 20 rare users then explain why some never come.

Quantitative vs qualitative vs mixed methods: comparison table

The three approaches differ in the question they answer, the data they collect, how they choose participants, how they analyze and how quality is judged. Creswell and Creswell's Research Design is the standard textbook comparison of the three [7].

How does sampling differ?

Quantitative studies aim for a sample that represents a population and is large enough to detect the effect of interest; qualitative studies choose people who can best inform the question, and stop when more data adds little.

For quantitative work, the sampling method decides how far you can generalize. A simple random or stratified sample supports claims about the population; a convenience sample does not. Our guide to probability vs non-probability sampling covers each method. The number comes from a power calculation: see statistical power and sample size.

For qualitative work, purposive sampling picks participants with the experience you study, sometimes chosen to vary on key traits (maximum variation). Grounded theory uses theoretical sampling: the next participant is chosen to test or fill out the emerging theory. On size, Guest, Bunce and Johnson found that in 60 interviews with a fairly similar group of women, most of the codes they would ever find had appeared within the first 12 [3]. Malterud, Siersma and Guassora's information power model is a more flexible guide: you need fewer participants when the aim is narrow, the sample highly specific, theory is used and the interviews are rich [4].

Example (illustrative), engineering: a study of how site engineers decide when to stop a concrete pour in bad weather interviews 14 engineers chosen from large and small firms and from both precast and in-situ work, stopping when new interviews stop adding new decision rules.

How does analysis differ?

Quantitative analysis summarizes and tests numbers with statistics; qualitative analysis codes data into themes or categories and interprets them.

For choosing a test, see which statistical test to use. For coding, see how to code qualitative data; reflexive thematic analysis as set out by Braun and Clarke is the most common approach [5].

What are the main mixed methods designs?

Creswell and Plano Clark's third edition sets out three core designs: convergent, explanatory sequential and exploratory sequential [6]. Earlier editions listed more designs (such as embedded and multiphase), which the third edition treats as complex designs built on the three core ones.

A common shorthand writes the strands in capitals for the priority strand, with an arrow for sequence and a plus for "at the same time": QUAN → qual, or QUAL + QUAN.

Convergent design

Collect quantitative and qualitative data in the same phase, analyze each separately, then merge the results to compare them.

• Use it when you want two views of the same question and have the time and skills to run both at once.
• Example (illustrative), health services: in one month, 300 patients rate a new online booking system on a satisfaction scale, while 20 patients and 10 receptionists are interviewed. A joint display shows where the ratings and the themes agree or conflict.
• Watch for: what you will do if the two sets of results disagree. Plan to explain it, not hide it.

Explanatory sequential design

Run a quantitative phase first, then a qualitative phase to explain its results, often choosing interviewees from the survey.

• Use it when you expect numbers to need explaining: a surprising result, outliers, or differences between groups.
• Example (illustrative), education: a survey of 900 secondary teachers finds that those in rural schools use a new digital platform much less. The researcher then interviews 16 rural teachers, chosen from high and low users, to find out why.
• Watch for: the time needed between phases, and deciding which results to follow up.

Exploratory sequential design

Run a qualitative phase first, then use its findings to build something (a survey, a scale, an intervention) that a quantitative phase tests.

• Use it when no good instrument exists for your population or concept, or you need to know what matters before you measure it.
• Example (illustrative), social science: focus groups with gig-economy delivery riders identify seven sources of work stress. These become items in a new questionnaire, which is piloted with 350 riders and checked for reliability and validity.
• Watch for: the instrument-building step takes longer than most students plan for. See reliability vs validity.

How do you choose between quantitative, qualitative and mixed methods?

Choose from the question first, then check feasibility: quantitative if you need to count, compare or test; qualitative if you need to explore, explain or understand experience; mixed methods if your question has both parts and you have time to integrate them.

Then check three things:

1. Time. Sequential designs need two phases of ethics, recruitment and analysis. A one-year master's often cannot fit both.
2. Skills. You need to do both statistics and qualitative analysis well. Supervisors can tell when one strand is an afterthought.
3. Word count. Mixed methods theses run longer. Check the limit before you commit.

How do you justify your choice in a methodology chapter?

State the approach, link it to your research question, explain what it lets you claim, and name the alternative you rejected and why. Cite a methods text and one or two studies in your field that used the same approach.

Example justification (illustrative): "This study uses an explanatory sequential mixed methods design [6]. The first research question asks whether rural and urban teachers differ in their use of the platform, which requires a survey large enough to compare groups. The second asks why any difference exists, which a survey cannot answer in depth. Interviews with teachers selected from the survey results allow the qualitative phase to explain the quantitative findings. A convergent design was rejected because the interview sample could not be chosen until the survey results were known."

The methodology chapter guide shows where this paragraph goes and what follows it.

Common mistakes

• Choosing qualitative because "I am not good at statistics". Qualitative analysis is just as demanding. Choose from the question.
• Calling a survey with two open questions "mixed methods". Without a plan to integrate the strands, it is a survey.
• Reporting qualitative findings as percentages. "65% of participants said…" from 20 interviews suggests a precision the design does not have.
• Using the wrong quality criteria. Do not judge interviews by statistical power, or a survey by saturation.
• Leaving integration to the discussion. Say in your methods where and how the strands meet.

How to choose your research method with AI for Research

AI for Research helps you see how others studied your topic and draft a plan for your own.

1. Find studies by approach (0 credits). In Find Papers, filter by study type, for example qualitative, cross-sectional or RCT, to see which approaches dominate your topic and where the gaps are.
2. Draft the methodology (3 credits). The Methodology builder lets you choose quantitative, qualitative or mixed methods, or ask it to suggest one, and drafts the design, sampling and analysis with citations to studies its search found.
3. Analyze the numbers (free). The data analysis tool runs descriptives, correlations, t-tests and chi-square in your browser; your file is not uploaded.
4. Draft the proposal. The research proposal generator has a method setting (quantitative, qualitative, mixed or suggest one) and runs 3 times a day without an account.

The Free plan has 45 credits a month; see pricing. Limits: the tools cannot know your access to participants or your program's rules, and methods textbooks are often missing from search results, so add them yourself.

FAQ

What are the three types of research methods?

The three broad types are quantitative, qualitative and mixed methods. Quantitative research collects numbers and tests relationships or differences with statistics. Qualitative research collects words, images or observations and interprets meaning, experience and process. Mixed methods research collects both and integrates them in one study, so that each strand helps explain, test or build on the other. Your research question decides which you need.

What is the main difference between qualitative and quantitative research?

Quantitative research asks how much, how many or how strongly, and answers with numbers from structured measures and larger samples. Qualitative research asks how, why or what it is like, and answers with themes or theory built from open-ended data from smaller, purposively chosen samples. The difference is in the question, the data and the kind of claim you can make, not in which is more scientific.

Is a survey quantitative or qualitative?

A survey with closed questions, such as rating scales and multiple choice, is quantitative because the answers are counted and analyzed statistically. A survey with only open-ended questions produces qualitative data. Many surveys mix both. If you analyze the open answers with thematic analysis and integrate them with the numbers, you may have a small mixed methods study, but only if the integration is planned.

What are the three core mixed methods designs?

Creswell and Plano Clark describe three core designs. In a convergent design, quantitative and qualitative data are collected in the same phase and merged. In an explanatory sequential design, a quantitative phase comes first and a qualitative phase explains its results. In an exploratory sequential design, a qualitative phase comes first and its findings shape a quantitative phase, such as a new survey instrument.

How many participants do I need for qualitative research?

There is no fixed number. Sample size in qualitative research is judged by whether you have enough data to answer the question in depth. Guest and colleagues found most themes appeared within the first 12 interviews in a fairly homogeneous sample. Malterud and colleagues' information power model says you need fewer participants when the aim is narrow, the sample is specific and the dialogue is rich.

Next steps

• Pick the specific design: types of research design.
• Choose who to study: probability vs non-probability sampling.
• Check your measures: reliability vs validity.
• Analyze interviews: how to code qualitative data.
• Write it up: how to write a methodology chapter.

Sources

1. Johnson, R. B., Onwuegbuzie, A. J., & Turner, L. A. (2007). Toward a definition of mixed methods research. Journal of Mixed Methods Research, 1(2), 112–133. https://doi.org/10.1177/1558689806298224
2. Fetters, M. D., Curry, L. A., & Creswell, J. W. (2013). Achieving integration in mixed methods designs: principles and practices. Health Services Research, 48(6 Pt 2), 2134–2156. https://doi.org/10.1111/1475-6773.12117
3. 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
4. Malterud, K., Siersma, V. D., & Guassora, A. D. (2016). Sample size in qualitative interview studies: guided by information power. Qualitative Health Research, 26(13), 1753–1760. https://doi.org/10.1177/1049732315617444
5. Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa
6. Creswell, J. W., & Plano Clark, V. L. (2018). Designing and Conducting Mixed Methods Research (3rd ed.). SAGE.
7. Creswell, J. W., & Creswell, J. D. (2018). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches (5th ed.). SAGE.