Qualitative vs quantitative research: key differences, with examples

Quantitative research collects numerical data and analyzes it with statistics to measure how much, how often or how strongly things are related, usually across a large sample. Qualitative research collects words, images and observations, from interviews, fieldwork or documents, and interprets them to understand how people experience and explain something. In short, quantitative research counts and tests, and qualitative research describes and interprets. A study that combines the two on purpose uses mixed methods.

Below: one topic studied both ways, the differences side by side, paired questions from ten fields, how to choose, and a short quiz.

One topic, studied both ways

The clearest way to see the difference is one topic studied twice. Both studies are invented, so the numbers show the form of a result, not a finding.

The first study tells the library how big the difference is, but not why. The second finds a reason the first never measured but can't say how common it is. Planned so that each informs the other, they would make a mixed-methods study.

Notice what the two labels describe. Numbers versus words is the data type. How a study is arranged to answer its question, as an experiment, a survey, a case study or an interview study, is its design. The two don't map one to one: a survey can collect both kinds of data, an experiment can add interviews, and a case study can rest on sales records.

The differences at a glance

What is quantitative research?

Quantitative research answers questions with numbers. The American Psychological Association's reporting standards send researchers to the quantitative standard when they collect their data as numbers or report them through statistical analyses. The numbers come from instruments you design, such as a closed-question survey or a test, or from existing records such as admissions, sales or census tables. The designs that use them fall into four families:

• Descriptive: how common something is, such as how many students work over 20 hours a week.
• Correlational: whether two variables move together, such as paid work and grades, without changing either.
• Quasi-experimental: comparing groups you did not assign, such as two schools before and after one changes its timetable.
• Experimental: randomly assigning participants to conditions, as a randomized controlled trial does, so a difference in outcomes can be traced to the condition.

Experiments give the most direct evidence about cause, because random assignment balances other differences between the groups on average. A quasi-experiment can support a causal claim when its comparison rules out the main alternative explanations; descriptive and correlational designs, on their own, usually can't. The prediction is usually stated in advance as a hypothesis and tested against a null hypothesis: see the hypothesis examples and the null hypothesis examples.

What is qualitative research?

Qualitative research answers questions with language and observation. APA's qualitative standard applies “when you collect your study data in the form of natural language and expression”: interviews, focus groups, field notes, documents and images. The researcher codes that material, groups the codes into themes, and builds an interpretation that stays close to the evidence.

Samples are chosen on purpose rather than at random: people who have been through the experience you are studying, or settings where it happens. Size is not set by a formula. Many studies keep collecting until new data stop adding anything, a criterion Barney Glaser and Anselm Strauss called “theoretical saturation” in The Discovery of Grounded Theory (1967). Benjamin Saunders and colleagues, writing in Quality & Quantity in 2018, identified four distinct approaches to saturation in current use, so say which one you mean.

The main traditions are phenomenology (what an experience is like), grounded theory (building theory from the data), ethnography (long-term fieldwork in a group), the case study and narrative research. Quality is judged differently too. Instead of reliability coefficients and statistical generalization, qualitative researchers show their trail from data to findings, state their own position, and describe the setting in enough detail that readers can judge whether the findings transfer to theirs.

Examples by field

Each row asks about one topic both ways. The quantitative question names variables and a comparison; the qualitative one names a group and an experience.

For more questions in each form, and how to narrow one, see the research question examples.

Qualitative vs quantitative data

The data, not the topic, decides which kind you have, and the cases in the middle trip people up. Whether you collected it yourself is a separate question, covered in primary vs secondary sources.

How to choose between them

1. Start from the question

   How many, how much, how often, or does one thing change another: you need numbers. How people experience, understand or go about something: you need qualitative data. Both: consider mixed methods.

2. Check what is already known

   A new topic with no established measures usually needs qualitative work first, to learn what to measure. A topic with validated scales and earlier studies is ready for testing.

3. Check access, time and skills

   Hundreds of respondents or a large dataset point one way; a few people willing to talk at length point the other. Statistics need training and a large enough sample, and interviews need transcription and slow coding.

4. Match your field

   Read recent studies in the journal or department you are writing for, and follow their conventions.

Whatever you choose, the method section has to justify it. The methodology examples show a quantitative experiment and a qualitative interview study written up in full, plus a shorter mixed-methods sample.

Mixed methods: using both

An NIH report on mixed methods in the health sciences, by John Creswell and colleagues (2011), pairs quantitative research on how large and how frequent something is with qualitative research on what it means, and stresses “intentionally integrating or combining these methods.” The basic designs differ in timing:

• Convergent: collect both kinds of data at about the same time and merge the results, as when a survey and interviews run in the same semester.
• Explanatory sequential: numbers first, then interviews to explain them, such as talking to card holders at the late-opening branches that gained the fewest visits.
• Exploratory sequential: qualitative work first, then a survey or instrument built from what it found.

The test is integration: a section or table where findings from each strand are set side by side and interpreted together.

Common mistakes and myths

The costliest mistake is a mismatch: “How do teachers experience the new curriculum?” answered with a five-item rating scale, or a causal question answered with six interviews.

Once the methods section is drafted, you can ask the Strokes Assistant to review it against your research question; it points out what is unclear or unsupported, such as a causal word in a correlational design.

Test yourself: qualitative or quantitative?

Decide for each study, then check the answers below.

1. A researcher records how many minutes 60 toddlers spend with each of four toys.
2. A historian reads 300 letters emigrants sent home to see how they described their new country.
3. A poll asks 2,000 voters to rate their trust in the courts from 1 to 7.
4. Eight focus groups of bus drivers talk about working night routes.
5. A hospital compares infection rates before and after a new hand-washing routine.
6. A team sorts 500 written customer complaints into categories and reports how often each appears.
7. An anthropologist lives in a fishing village for a year and keeps field notes.
8. A newsletter sends two subject lines to 10,000 readers each and compares open rates.