Choosing Between Words and Numbers: A Practical Guide to Your Research Approach

Roughly speaking, every empirical thesis answers one of two questions: how many? or why? Quantitative research counts, measures, and compares; qualitative research explores meaning, experience, and context. Many students lose weeks because they pick a method before they have settled their research question, then try to force the two together. The order should always be the reverse: your question decides your method, not the other way round.

This guide explains what each approach is good at, when to combine them, and how to make a confident, defensible choice for your project.

What quantitative research does well

Quantitative research works with numerical data and statistical analysis. You define variables in advance, collect structured data (surveys with scales, experiments, existing datasets), and test hypotheses to find patterns that generalise to a wider population.

Choose a quantitative design when you want to:

• Measure how often something happens or how strongly two things relate.
• Compare groups or conditions.
• Test a hypothesis you can state before collecting data.
• Generalise from a sample to a larger population.

The pay-off is precision and breadth. The trade-off is depth: numbers tell you that a relationship exists, rarely why it exists in the lived experience of your participants.

Tip from practice: If you can write your research question as “Is there a relationship between X and Y?” or “Does group A differ from group B?”, you are almost certainly looking at a quantitative design.

What qualitative research does well

Qualitative research works with non-numerical data: interviews, focus groups, open-ended responses, documents, observations. Instead of testing a fixed hypothesis, you build understanding inductively, letting themes emerge from the material.

Choose a qualitative design when you want to:

• Explore a topic that is poorly understood or under-researched.
• Understand motivations, perceptions, and meaning.
• Describe a process or experience in rich detail.
• Generate new theory or hypotheses rather than confirm existing ones.

The pay-off is depth and nuance. The trade-off is reach: a study of twelve interviews gives you a vivid picture, but you cannot claim it represents an entire population.

A simple decision rule

In our coaching practice we often see students agonise over this choice when their own question already contains the answer. Try this rule of thumb, the Question–Data–Claim test:

1. Question — Are you asking “how much / how many / does X predict Y” (quantitative) or “how / why / what is it like” (qualitative)?
2. Data — Can the information you need realistically be turned into numbers, or does it live in language, behaviour, and context?
3. Claim — Do you want to generalise to a population (quantitative) or understand a case in depth (qualitative)?

If all three answers point the same way, your method is clear. When they conflict, that tension is usually a sign your research question still needs sharpening, which is exactly the kind of thing worth discussing early with a coach or supervisor.

When to combine both: mixed methods

You do not always have to choose one camp. A mixed-methods design uses both, and it is genuinely powerful when each part answers something the other cannot. For example, you might run a survey to measure how widespread an attitude is, then interview a subset of respondents to understand the reasoning behind it.

Mixed methods are demanding, though. You are effectively running two studies and then integrating them, which doubles the design, data-collection, and analysis work. For a bachelor’s project with a tight timeline, that is often too much. For a master’s thesis or dissertation, it can be the right level of ambition, provided you plan it deliberately rather than bolting interviews onto a finished survey.

Matching method to feasibility

The “best” method on paper is worthless if you cannot execute it in the time and with the access you have. Before you commit, sanity-check three practical questions:

• Access — Can you actually reach enough participants? Two hundred survey responses and twenty interviews are very different recruiting challenges.
• Skills — Are you comfortable with statistical analysis, or with systematic coding of text? Both are learnable, but they take time.
• Scope — Does the method fit the word count and deadline of your degree?

Remember: There is no inherently “more scientific” method. A poorly designed quantitative study is no better than a poorly designed qualitative one. Rigour comes from a method that genuinely fits the question.

Turning your choice into a plan

Once you have chosen, write down the consequences. A quantitative path means you should plan your sampling, your measurement instruments, and the statistical tests you will run, ideally before you collect a single data point. A qualitative path means planning your interview guide, your sampling logic, and your coding strategy. Either way, this thinking belongs in your research proposal, where your supervisor can flag problems while they are still cheap to fix.

If you are still unsure, this is one of the most valuable moments to get an outside perspective. Our statistics support coaching helps you pressure-test your design, choose an approach you can defend, and avoid the costly mistake of changing methods halfway through. The decision you make here shapes everything that follows, so it is worth getting right before the work begins.