Class_generator/DIAGNOSTIC_HOMEWORK_PLAN.md

# Diagnostic Homework System - Project Plan

## 🎯 OBJECTIVE
Create diagnostic homework to assess student readiness for English exams in China, then build personalized exam prep plans.

---

## 📊 CURRENT SITUATION

### Students
- **Two groups**: SBS2 level & SBS8 level (mixed class)
- **Last lesson**: Chapter 2-8-9 fusion (combined content)
- **Next goal**: Prepare for English exams (Chinese exam format)

### Exam Format (Chinese English Exams)
- **NOT production-heavy** → Recognition-focused
- **Multiple choice** (vocab, grammar)
- **Fill-in-blanks** with options
- **Reading comprehension** with QCM
- **Error correction** (identify wrong form)
- **Limited open questions** (short answers)

### Problem
- Don't know **exact weak points** of each student
- Need **diagnostic data** before creating targeted prep plan
- Can't prepare effective revision without understanding current level

---

## 💡 SOLUTION: Diagnostic Homework

### Format (Simple & Effective)
**Each homework contains:**
1. **1 Text** (120-140 words, known vocabulary, complex sentences)
2. **3 Comprehension Questions** (test reading ability)
3. **1 Audio** (teacher-recorded, ~30-40 seconds)
4. **3 Production Questions** (open-ended, graduated difficulty)

### Two Separate Homeworks
- **Homework A**: SBS2 level (focus: locations, where questions, to be)
- **Homework B**: SBS8 level (focus: clothing, present continuous, demonstratives)

---

## 📋 HOMEWORK STRUCTURE DETAILS

### Text Requirements
- **Length**: 120-140 words (challenging but manageable)
- **Vocabulary**: ONLY from their chapters (cumulative: Ch1 + Ch2 for SBS2, Ch1-8 for SBS8)
- **Complexity**: Long sentences, multiple clauses, connectors
- **Goal**: Test comprehension, not vocabulary recognition

### 3 Comprehension Questions
**Test reading skills:**
1. **Literal**: Direct answer in text (Where is X? What color is Y?)
2. **Inferential**: Requires understanding context (Why is X happening?)
3. **Detail**: Tests careful reading (specific information retrieval)

**Format**: Open-ended short answers (not multiple choice - want to see their writing)

### Audio Component
- **Teacher records** custom audio (~30-40 seconds)
- **Content**: Related to chapter theme but independent from text
- **Purpose**: Diagnostic for listening comprehension (analyzed through written production answers)

### 3 Production Questions (Graduated Difficulty)

**Q1 - Linked to Text** (直接相关)
- Application of text content
- Uses same grammar/vocab from reading
- Example SBS2: "Where is [person from text]?"
- Example SBS8: "What is [person from text] wearing? Why?"

**Q2 - Semi-Linked** (半相关)
- Same topic/grammar but new context
- Requires vocabulary extension
- Example SBS2: "Where do you go on weekends? Why?"
- Example SBS8: "What do you wear when it's very cold? Describe."

**Q3 - Creative/Free** (创意开放)
- **CRITICAL**: Tests linguistic flexibility & creativity
- Forces students to adapt vocabulary to unexpected context
- **NOT predictable** (avoid "What are you wearing today?")
- Reveals true production capability vs memorization

**Q3 Examples:**
- SBS2: "Do pets have special places at home? Where? Describe."
- SBS8: "Do dogs/cats wear clothes in your country? If yes, what? If no, why not?"
- Alternative SBS8: "Design a school uniform - what colors, what clothes?"

---

## 🎯 DIAGNOSTIC DATA COLLECTED

### Reading Comprehension
- Can they understand complex sentences with known vocabulary?
- Literal vs inferential comprehension level
- Speed/accuracy of information retrieval

### Listening Comprehension
- Analyzed indirectly through production answers (no direct listening questions)
- Teacher observes if they understood audio context

### Grammar Accuracy
**SBS2 Focus:**
- Where questions formation
- To be (am/is/are) usage
- Prepositions (at/in/on)
- Subject pronouns

**SBS8 Focus:**
- Present continuous (am/is/are + -ing)
- This/that/these/those
- Singular/plural agreement
- Adjective placement

### Production Capability
**3 levels revealed:**
1. **Reproduction**: Can they reuse text structures? (Q1)
2. **Extension**: Can they apply to new contexts? (Q2)
3. **Creation**: Can they produce flexibly and creatively? (Q3)

### Vocabulary Mastery
- Active vs passive vocabulary
- Correct usage in context
- Spelling accuracy

---

## 📈 NEXT STEPS (After Diagnostic)

### Phase 1: Analysis
- Collect homework responses
- **AI-powered analysis** (LLM validates & scores)
- Generate individual diagnostic reports:
  - Vocab gaps
  - Grammar weak points
  - Comprehension level
  - Production capability score

### Phase 2: Personalized Exam Prep Plan
Based on diagnostic data, create:

**For Each Student:**
- **Targeted drill exercises** (focus on weak grammar points)
- **Vocabulary reinforcement** (words they missed/misused)
- **Question-type practice** (Chinese exam format)

**Study Plan Structure:**
1. **Diagnostic Session** (1 homework) ← WE ARE HERE
2. **Drill Phase** (3-4 sessions, targeted exercises)
3. **Mock Exam Phase** (2 sessions, real format practice)
4. **Review & Polish** (1 session, confidence building)

### Phase 3: Exam Prep Modules

**Module Types to Build:**

**A. Drill Exercises** (Grammar Pattern Recognition)
- Multiple choice (do/does, is/are, this/these)
- Fill-in-blanks with options
- Error correction practice
- Speed drills (timed)

**B. Vocabulary Recognition**
- Matching exercises (word ↔ Chinese)
- Synonym/antonym recognition
- Word-in-context (choose correct usage)

**C. Reading Comprehension**
- Short texts + QCM questions
- Chinese exam format
- Progressive difficulty

**D. Mock Exams**
- Full Chinese exam format
- Strict timing
- Auto-scoring
- Error analysis with AI feedback

### Phase 4: AI Integration

**AI Features:**
- Auto-scoring of open questions
- Pattern recognition (common errors)
- Personalized feedback in Chinese + English
- Adaptive difficulty (unlock harder exercises after mastery)
- Progress tracking & visualization

---

## 🛠️ TECHNICAL IMPLEMENTATION

### Current System (DRS)
- Modular architecture (VocabularyModule, TextAnalysis, etc.)
- AI-powered scoring (DeepSeek API)
- Content JSON structure
- TTS for audio (can use for text, but teacher records custom audio)

### To Build
1. **Diagnostic Homework Module**
   - Text display
   - Audio playback (teacher upload)
   - Open question input
   - Submit & store responses

2. **AI Analysis Engine**
   - Score comprehension answers (0-100)
   - Grammar error detection
   - Vocabulary usage analysis
   - Generate diagnostic report

3. **Exam Prep Exercise Generator**
   - Pull from chapter content
   - Generate targeted drills based on weak points
   - Progressive unlocking system

4. **Mock Exam Module**
   - Chinese exam format template
   - Timer integration
   - QCM + fill-in-blanks + error correction
   - Auto-scoring + detailed feedback

---

## 📝 IMMEDIATE ACTION ITEMS

### Now (Current Task)
- [x] Define homework structure
- [x] Clarify diagnostic goals
- [ ] **Create SBS2 homework** (text + 6 questions)
- [ ] **Create SBS8 homework** (text + 6 questions)
- [ ] Teacher records audio for both
- [ ] Students complete homework

### After Homework Collection
- [ ] Analyze responses
- [ ] Generate diagnostic reports
- [ ] Design personalized drill exercises
- [ ] Build exam prep modules

---

## 🎓 PEDAGOGICAL PRINCIPLES

### Why This Approach Works

**1. Diagnostic-First**
- No guessing student levels
- Data-driven prep plan
- Efficient use of limited time before exam

**2. Graduated Difficulty**
- Q1 (linked) → builds confidence
- Q2 (semi-linked) → tests application
- Q3 (creative) → reveals true ability

**3. Exam-Format Alignment**
- Chinese exams = recognition > production
- Drills match actual exam question types
- Mock exams = confidence building

**4. Personalization**
- Each student gets targeted weak-point drills
- No wasted time on mastered content
- Efficient exam prep

**5. AI-Powered Scaling**
- Teacher time = diagnostic design + audio recording
- AI handles scoring, analysis, feedback
- Scalable to many students

---

## 📊 SUCCESS METRICS

### Diagnostic Phase
- ✅ 100% homework completion rate
- ✅ Clear identification of weak points per student
- ✅ Actionable data for prep plan

### Prep Phase
- ✅ Students improve on targeted weak points
- ✅ Mock exam scores increase session-to-session
- ✅ Confidence levels rise

### Exam Results
- ✅ Students score at or above expected level
- ✅ Fewer "silly mistakes" (grammar/vocab errors)
- ✅ Strong comprehension question performance

---

## 🔑 KEY INSIGHTS

### What Makes This Different
**Traditional homework**: Generic exercises, no diagnostic value
**This system**: Strategic diagnostic → personalized prep → exam success

### Critical Success Factors
1. **Q3 creativity** = reveals true linguistic flexibility
2. **Text complexity** = tests comprehension vs vocab memorization
3. **AI scoring** = scalable, consistent, immediate feedback
4. **Exam format alignment** = practice = performance

### Future Expansion
- Scale to more students (same diagnostic approach)
- Different exam types (adapt question format)
- Automated homework generation (AI creates texts from chapter vocab)
- Progress tracking dashboard (student + teacher views)

---

## 📌 SUMMARY (TL;DR)

**Problem**: Need to prep students for Chinese English exams, don't know their exact weak points

**Solution**: Diagnostic homework (1 text, 3 comp Q's, 1 audio, 3 prod Q's) → reveals gaps

**Next**: Build personalized exam prep (drills, mock exams, AI feedback) based on diagnostic data

**Goal**: Data-driven, efficient, targeted exam preparation that matches Chinese exam format

**Status**: Designing diagnostic homework content now → implementation → student testing → analysis → prep plan

---

**END OF DOCUMENT**