Picking the wrong research methodology is the most common mistake at the proposal stage. Students often pick based on what they're comfortable with rather than on what their question actually requires. Here's a clear guide to the five main approaches — what each is for, what each is bad at, and how to know which one your specific topic needs.
1. Quantitative Research
What it is: Numbers. Statistical analysis. Surveys with closed-ended questions. Experiments with measurable outcomes.
Best for: Questions where you can measure something. "Does X cause Y?" "How much does A differ from B?" "Is there a relationship between X and Y in this population?"
Worst for: Questions about meaning, experience, or context. You can't survey "what does poverty feel like" — you can survey "how many days last month did you skip meals," which is a different question.
2. Qualitative Research
What it is: Words. Interviews, focus groups, ethnographic observation. Open-ended questions. Thematic analysis of texts.
Best for: Questions about experience, meaning, social processes. "What is it like to X?" "How do people make sense of Y?" "What practices emerge in this community?"
Worst for: Questions of generalisability or causation. You can't conclude that "X causes Y in the general population" from 12 interviews.
3. Mixed Methods
What it is: Both. Usually quantitative gives you the breadth, qualitative gives you the depth. Done well, the two illuminate each other.
Best for: Questions that need both — "How widespread is X, and what does it look like in practice?" Often the strongest dissertation methodology if you can pull it off.
Worst for: Students with limited time. You're effectively doing two studies. Budget accordingly.
4. Case Study
What it is: Deep examination of one or a small number of cases. Can use any data type — interviews, documents, observation, statistics.
Best for: Questions about complex phenomena in real-world contexts. "How did organisation X navigate this transition?" "What happened in this specific event and what does it suggest?"
Worst for: Generalisation. Case studies illuminate the specific case; they make claims about the general only with caution.
5. Systematic Literature Review
What it is: A structured analysis of existing published research, using explicit inclusion/exclusion criteria and rigorous synthesis methods.
Best for: Questions about the state of the research literature, gaps in existing work, or aggregating findings across many studies. Sometimes a complete dissertation; often a chapter within one.
Worst for: Topics where the literature is thin or where original empirical work is what your discipline expects.
How to pick yours
Three questions, in order:
1. What is the precise question?
Not "I want to study X." The actual research question, in one sentence, with the key variables specified. The question dictates the methodology.
2. What kind of evidence would answer it?
Numbers? Words? Both? A specific case? An aggregation of existing work? The answer to "what evidence is needed" tells you which methodology is appropriate.
3. What are you actually able to do?
If your methodology requires 200 survey respondents and you have access to 30, you have a different methodology problem than you thought. Pick the methodology you can actually execute.
The mistakes to avoid
- Picking quantitative because it "feels more rigorous." It's only more rigorous when it's the right method for your question.
- Picking qualitative because it "seems easier." It's usually harder. Doing thematic analysis well is genuinely demanding.
- Picking mixed methods because it "covers all bases." Only choose mixed methods when both methods serve specific aspects of your question. Otherwise you're doing two mediocre studies.
- Picking the methodology before the question. Methodology serves the question. Get the question right first.
One small habit that helps
Write your research question at the top of every methodology document you produce. Every time. The question keeps you honest. The number of times I've watched a student's methodology drift away from their question over six months is large enough to suggest this small habit prevents real damage.
Try this
Write your research question in one sentence. Now ask: what evidence would answer this question? Now ask: which of the five methodologies above produces that evidence?
The answer should be obvious within five minutes. If it isn't, the question needs sharpening before the methodology decision can be made.
