Choosing between qualitative, quantitative and mixed methods is not a matter of preference. It follows from what you want to find out, what data you can realistically collect, and who or what you can reach — and when those three point different ways, that conflict is telling you the question needs adjusting.
Answer three questions and this tool suggests an approach, a design, a collection method and an analysis strategy, showing its reasoning and its caveats every time. There is no universally correct methodology for a topic, and any tool that implies otherwise is misleading you.
Quantitative research measures and counts. It can establish how common something is, whether groups differ, and whether variables move together — and with the right design, whether one causes another. It cannot tell you what any of it means to the people involved.
Qualitative research works with accounts in participants' own words. It can show how something is understood and experienced, and surface factors nobody thought to put on a questionnaire. It cannot say how widespread anything is and does not generalise statistically. Mixed methods does both at close to double the workload, which is why one strand should lead.
Suppose your aim is to test whether one thing causes another but your data will be interviews. Those are incompatible: accounts cannot establish causation. The honest fixes are to find a measurable indicator or to reframe the question as exploratory — not to count quotes and present the tally as statistics.
The opposite conflict is equally common: an exploratory aim with numeric data. Numbers cannot tell you how something is experienced. This tool flags both cases rather than smoothing over them, because the conflict is real information about your design.
Below roughly thirty participants most statistical tests cannot detect anything reliably — you may find nothing purely because the sample was small, which tells you nothing about the world. For qualitative work, eight to fifteen interviews is a normal range and the goal is variety, not volume.
This is why access constrains design so heavily, and why fieldwork logistics matter in Nepal: travel time to a district, the season, and how many households you can realistically visit all shape what design is honest. If you can reach twelve people, a qualitative design is the defensible choice — a twelve-person survey reporting percentages invites immediate criticism.
The design is the logic: cross-sectional, comparative, correlational, quasi-experimental, case study, evaluation. The method is how data arrives: questionnaire, interview, focus group, observation, secondary records. The analysis is what you do with it: descriptive statistics, tests of difference, regression, thematic analysis.
Students often name a method and call it a methodology. "I am doing a questionnaire" describes the method alone. A methodology chapter needs all three plus the reasoning tying them to the research question — which is what this tool lays out.
Conventions differ by discipline. A design that reads as rigorous in public health can look thin in sociology, and the reverse. Departments have preferences and external examiners have expectations that are rarely written down.
Treat this output as a structured starting point for a supervision meeting: a defensible option with its reasoning attached is a far better basis for discussion than an open question. Your supervisor knows what your examiners expect, and no tool can substitute for that.
No. Several designs can be defensible for the same topic and conventions vary by discipline. What matters is that the design fits your question and that you can justify it.
Only if you genuinely need both kinds of evidence and have time for both. Mixed methods roughly doubles the work, and a weak second strand does more harm than good.
For statistical tests, generally at least thirty and often many more. For interview-based work, eight to fifteen chosen purposively for variety is common.
No. It can show how participants explain a process and suggest mechanisms worth testing, but a causal claim needs a design that compares conditions.
The method is the technique — questionnaire, interview, observation. The methodology is the whole logic: design, method, analysis and the reasoning connecting them to your research question.