# Promise Precision and Venture Funding **🔗 작전 지휘부**: `../../삼도수군/` (일일 순환 추적) **🏛️ 전투 본부**: 이 폴더 (최종 결과물) --- ## 1. Hypotheses **H1 (Early Stage - Short-term Cost):** α₁ < 0 😵‍💫 Vague promises hurt at first → Lower initial funding **H2 (Later Stage - Long-term Benefit):** β₁ > 0 😵‍💫 Vague promises benefit later → Higher survival probability --- ## 2. Data ### Source Longitudinal data with company descriptions and funding data at early and later stages from venture databases (Pitchbook, Crunchbase, or similar). ### Sample Structure * **Temporal**: Early-stage investments → Later-stage outcomes * **Cross-section**: Technology ventures with detailed descriptions * **Panel**: Same firms tracked across funding rounds * **Target N**: ~30-75 firms with complete data ### Key Variables | Variable | Measurement | Role | |----------|-------------|------| | 😵‍💫 Vagueness | Inverse of linguistic certitude (LIWC) in company description | Independent | | 💰 Early_Funding | Funding amount at early stage ($ millions) | Dependent (Model 1) / Control (Model 2) | | 🎯 Later_Success | Binary: 1 if raised later-stage funding, 0 otherwise | Dependent (Model 2) | | Team_Size | Number of employees at early stage | Control | | Prior_Exit | Founder has previous exit (binary) | Control | | Sector | Industry category | Control | --- ## 3. Models ### Model 1: Early-Stage Funding Amount (OLS) ``` 💰 Early_Funding_i = α₀ + α₁ · 😵‍💫Vagueness_i + Controls + ε_i ``` **Expected:** α₁ < 0 (vagueness → less initial funding) ### Model 2: Later-Stage Success (Logistic) ``` logit(🎯 Later_Success_i) = β₀ + β₁ · 😵‍💫Vagueness_i + β₂ · 💰Early_Funding_i + Controls + ε_i ``` **Expected:** β₁ > 0 (vagueness → better survival, controlling for early funding) --- ## 4. Expected Patterns ### Pattern A: Early Stage (Short-term Cost) ``` Funding ↑ | |\ | \ | \_____ |____________ 😵‍💫 Vagueness → Precise Vague ``` **Interpretation:** Vague descriptions → lower early-stage funding ### Pattern B: Later Stage (Long-term Benefit) ``` Success ↑ | ____/ | / | / | / |/____________ 😵‍💫 Vagueness → Precise Vague ``` **Interpretation:** Vague descriptions → higher later-stage success rate --- ## 5. Analysis Pipeline ### 파이프라인 위치 ``` code/ ├── 01_process_company_data.py (데이터 추출 & 정제) ├── 02_process_deal_data.py (펀딩 데이터 처리) ├── 03_create_panel.py (패널 구성) ├── 04_run_analysis.py (Model 1 & 2) └── 05_create_deliverables.py (Tables & Figures) ``` ### 단계 1. **Data Prep**: Extract funding records, compute vagueness scores 2. **Model 1**: OLS regression (early funding ~ vagueness) 3. **Model 2**: Logistic regression (later success ~ vagueness + early funding) 4. **Visualization**: Create Pattern A & B with actual data 5. **Robustness**: Sector effects, alternative vagueness measures --- ## 6. Deliverables | Item | 위치 | 설명 | |------|------|------| | **Table 1** | `output/table1_descriptive.csv` | Descriptive statistics | | **Table 2** | `output/table2_model1.csv` | Model 1 results (early stage) | | **Table 3** | `output/table3_model2.csv` | Model 2 results (later stage) | | **Figure 1** | `output/figure1_early.png` | Vagueness → Early funding | | **Figure 2** | `output/figure2_later.png` | Vagueness → Later success | --- ## 🔄 견리사의 순환 작업법 이 연구는 **삼도수군 폴더**에서 일일 순환으로 진행됩니다: ``` 아침 (見) ↓ 삼도수군/전투일지.md (계획) ↓ 작업 (利 → 思 → 義) ↓ 삼도수군/1_利/ (ChatGPT 프로토타입) 삼도수군/2_思/ (Claude 정교화) 삼도수군/3_義/ (Gemini 검증) ↓ ✅ 검증 완료 ↓ 이 폴더 (empirics/)로 이동 ↓ code/ (최종 코드) output/ (최종 결과) ``` **일일 워크플로우**: `../../삼도수군/README.md` 참조 --- ## 📅 3주 계획 ### Week 1 (10.25-10.31): Data + Model 1 **목표**: Table 1, 2 완성 - Day 1-2: 데이터 파이프라인 구축 - Day 3-4: Model 1 구현 - Day 5-7: 검증 & Table 2 **산출**: `output/table1*.csv`, `output/table2*.csv` --- ### Week 2 (11.01-11.07): Model 2 + Visualization **목표**: Table 3, Figure 1, 2 완성 - Day 8-10: Model 2 (Logistic) - Day 11-12: 시각화 - Day 13-14: Robustness checks **산출**: `output/table3*.csv`, `output/figure*.png` --- ### Week 3 (11.08-11.15): Paper **목표**: 논문 초고 - Day 15-17: Introduction, Theory, Method - Day 18-19: Results, Discussion - Day 20-21: 최종 제출 **산출**: `../theory/paper_draft.md` --- ## 🎯 핵심 연결점 ### 이론적 기반 - **OIL Framework**: τ* = max{0, √(V/4i) - 1} - **Strategic Ambiguity**: When should ventures be vague? - **Signaling Theory**: What do vague promises signal? ### Empirical Contribution Promise precision has **non-monotonic effects**: - **Short-term** (H1): Precision helps (credibility) → α₁ < 0 - **Long-term** (H2): Vagueness helps (flexibility) → β₁ > 0 --- ## 🚨 막힐 때 ### 연구 설계 질문 → 이 파일 재독 ### 일일 작업 막힘 → `../../삼도수군/README.md` ### 코드 문제 → `code/PIPELINE_GUIDE.md` --- **"신에게는 아직 12척의 배가 있습니다"** **이 workflow는 우리의 지도입니다** 🗺️