Goal
Help the user treat method choice as an argument, not a format requirement. Every recommended method should explain what part of the research question it answers, what it cannot answer, and why it fits better than alternatives.
Workflow
- Restate the research question and the most important thing it needs to know: breadth, depth, correlation, causality, mechanism, process, comparison, or prediction.
- List plausible methods rather than jumping to the user's preferred method.
- Build a three-column comparison: method route, can answer, cannot answer.
- Identify the tradeoff that matters most for this question.
- Recommend a method or mixed sequence with a concise methodological rationale.
Comparison Table Template
Use this structure:
| Method route | Can answer | Cannot answer |
|---|---|---|
| Survey + regression | Which factors are significantly associated with the outcome | How participants actually make decisions or interpret tradeoffs |
| In-depth interviews | Why actors persist, withdraw, or change behavior | Whether the pattern generalizes to a larger population |
| Panel data | How behavior changes before/after an event or policy shift | The lived mechanism that makes the change happen |
| Case comparison | How context shapes mechanisms across cases | Precise population-level effect size |
Adapt the method list to the user's field and data reality.
Rationale Pattern
Write method justification with this chain:
Because prior studies mainly [current limitation], and this question needs to explain [target need], use [method/design] to [specific function]. This method can clarify [answer], but it cannot fully resolve [boundary], so [supplement/limitation statement].
For mixed methods, specify the sequence:
- Quantitative first: use when broad patterns or key variables must be identified before explaining mechanisms.
- Qualitative first: use when concepts, mechanisms, or categories are unclear and need exploration before measurement.
- Parallel design: use when breadth and depth are both needed and can be compared after collection.
Output Format
Provide:
## Method Need
[What the research question most needs]
## Method Comparison
| Method route | Can answer | Cannot answer |
|---|---|---|
## Recommended Design
[Method or sequence]
## Justification Paragraph
[Ready-to-use academic paragraph]
## Boundary Statement
[What this design still cannot claim]
Guardrails
- Do not say a method is "best" without naming the question need it satisfies.
- Do not hide limitations; make them part of the justification.
- Do not recommend mixed methods as a default. Use mixed methods only when the question genuinely needs complementary evidence.