Research Foundations › Module 3: Types of research
Quantitative and qualitative research
Lesson 1 of 3 · 2 min
Research methods are often grouped by the kind of data they use.
Quantitative research
Works with numbers: measurements, counts, scores. It asks "how much?", "how many?", "is A bigger than B?". Most machine learning research is quantitative: accuracy, loss, speed, cost.
Strengths: results can be compared and tested with statistics. Weaknesses: numbers can hide why something happens, and a measure can miss what really matters.
Qualitative research
Works with words, images and observations: interviews, open-ended answers, field notes. It asks "why?", "how?", "what is it like?".
Strengths: it finds reasons and experiences that numbers miss. Weaknesses: it is harder to generalise from a small group, and the researcher's own views can colour the analysis, so the method must be described carefully.
| Quantitative | Qualitative | |
|---|---|---|
| Data | Numbers | Words, images, observations |
| Typical question | How much? Does A beat B? | Why? How do people experience this? |
| Example | Does a chatbot cut average reply time? | Why do some students stop using a study chatbot after a week? |
- Data
- Quantitative: Numbers
- Qualitative: Words, images, observations
- Typical question
- Quantitative: How much? Does A beat B?
- Qualitative: Why? How do people experience this?
- Example
- Quantitative: Does a chatbot cut average reply time?
- Qualitative: Why do some students stop using a study chatbot after a week?
Mixed methods
Many good studies use both. You might measure that 40% of users stop using an app after a week (quantitative), then interview some of them to learn why (qualitative). The numbers tell you that something happens. The words help explain it.
RememberNeither kind is "more scientific". Each answers a different type of question. Choose by the question, not by habit.