ueai-problem-framing

Problem framing (stage 00)

Inputs you receive

• {{TOPIC}} — short title
• {{QUESTION}} — one-sentence draft
• {{COURSE}} — 贸易数据库与分析工具 or 经济建模与写作

What you produce

A filled 00_problem/problem_brief.md with the following five fields:

1. 研究对象：what country / sector / firm / period
2. 核心变量：dependent + 2–4 candidate independent vars; clarify units
3. 可用数据：1–3 candidate databases with concrete table/field names
4. 可能方法：1–2 methods, each with the minimum assumption set
5. 结果展示：figure types, table types — be specific (e.g. "2×2 福利分解表" not "results table")

Plus a 可核验口径 checklist — 3–5 items the student must verify offline (HS code, country list, exchange rate base, deflator base year, etc.).

Decomposition heuristics

For trade-policy topics: object = countries × HS chapters × year window. Method = SMART / GTAP / partial-equilibrium / event-study. For digital-economy / FDI / GVC topics: identify a treatment or shock, identify a measurable outcome. For modelling-course topics: prefer a model the student can actually run (panel FE, DID, GTAP) over fancy ones (DSGE).

If the topic is too broad (e.g. "数字经济与中国发展"), refuse to scope and ask for one country + one period + one outcome.

Output template

# 问题拆解

项目题目：{{TOPIC}}
对应课程：{{COURSE}}
研究问题：{{QUESTION}}

## 本阶段任务
把宽泛兴趣转化为可研究、可检验、可表达的课程任务。

## AI辅助记录
- 使用时间：{{ISO timestamp}}
- 使用工具：{{model name}}
- 使用提示词：prompts/problem-framing.md
- AI反馈摘要：
  - 研究对象：...
  - 核心变量：...
  - 可用数据：...
  - 可能方法：...
  - 结果展示：...

## 可核验口径
- [ ] ...
- [ ] ...

## 人工核验与修改
- 数据来源或课程材料核验：
- 教师要求对照：
- 人工修改说明：

## 本阶段产出
- 提交文件：00_problem/problem_brief.md
- 课堂展示或讨论记录：
- 教师反馈：

What you DO NOT do

• Do not invent a "novelty" claim.
• Do not write a literature review here (that's stage 02 / 04).
• Do not pick the method definitively — propose 1–2 with tradeoffs.
• Do not fill ## 人工核验与修改. That's the student's column.